DEV Community: Mikhail Klimenko

How to put a Visio diagram into Confluence Cloud

Mikhail Klimenko — Mon, 05 Oct 2026 20:29:14 +0000

You attach the .vsdx, click it to check, and the browser downloads the file instead of showing the drawing. That is not a permissions problem or a broken upload. Confluence Cloud simply does not preview this format, and it is not going to start.

Short answer. Attaching the file works; previewing it does not. To get the diagram visible on a page you either export it from Visio as SVG and insert that, or import the .vsdx into a diagramming app inside Confluence and keep it editable there. Export to PNG only if the picture is small and finished forever. Which one you want depends on a single question: does the diagram change, and do its shapes carry links?

Why the attachment does not show

Confluence Cloud previews a long list of formats on the page. The modern Visio format is not on it: a .vsdx attachment downloads rather than opening. Atlassian has a ticket for exactly this, CONFCLOUD-68619, and the honest part is how it ended. It was raised in 2019, it says in plain words that there is no known workaround, and in October 2023 it was closed as timed out — not fixed, just expired. The older .vsd format did preview, which is why people with long memories insist this used to work.

So the decision is not whether to convert, but into what.

The four routes, and what each costs later

Export to PNG

The fastest and the most common. File, Export, PNG, drag onto the page, done in a minute.

The cost shows up later. A PNG is pixels, so the text inside the boxes turns to mush the moment somebody zooms to read a label — and architecture diagrams exist to be zoomed into. Any hyperlink you attached to a shape in Visio is gone, because pixels cannot hold a link. And every edit means exporting and re-uploading by hand.

Export to PDF

Sharper than PNG and it survives printing, which is why audit and compliance teams like it. But a PDF is a document sitting next to the page rather than a picture on it: the reader has to open an attachment, and the diagram stops being part of what they are reading. Fine as a formal record, poor as documentation.

Export to SVG

SVG is vector: it stays sharp at any zoom, the text stays text, and, crucially, the links attached to shapes travel inside the file. For a diagram with any text at all this is the right export format.

There is a catch, and it is the thing people write to us about: Confluence renders an uploaded SVG as a flat image. The file still contains everything, but the viewer does not make it live, so your clickable boxes arrive looking perfect and behaving like a screenshot.

Import into a diagramming app and keep it editable

If you want to stop depending on Visio altogether, import rather than export. The draw.io app for Confluence reads .vsdx: open a draw.io diagram, choose File then Import, pick the file. The result is an editable diagram living inside Confluence, so the next change is made where the documentation is, without a round trip through the desktop.

Two honest caveats. Complex drawings usually need small corrections after import, because Visio stencils and draw.io shapes are not a perfect match. And this is a swap of tools, not a conversion trick: your colleagues who live in Visio will still be working in Visio.

The arithmetic nobody does

Take one architecture diagram that gets a real edit once a month, which is modest for a system under active development.

By the export route each change is: open Visio, edit, export, find the Confluence page, delete the old attachment, upload the new one, check the page still looks right. Call it ten minutes if nothing goes wrong, and it goes wrong often enough — somebody uploads the picture but forgets the page still shows the old attachment. Two hours a year per diagram, for a file that in theory you already finished.

By the import route the edit happens on the page itself, and there is nothing to re-upload.

If the shapes have links on them

This is the case worth separating out. A Visio diagram where each box links to the runbook, the service page or the owner's documentation is not a picture: it is navigation. Flattening it to PNG destroys exactly the part that made it useful, and nobody notices until a reader clicks a box and nothing happens.

Export that diagram as SVG and the links survive in the file. Then they stop working in Confluence, for the viewer reason above.

Disclosure: I build the app that fixes this one. Clickable SVG renders the SVG as live markup rather than a flat picture, so the links your Visio shapes already carry keep working on the Confluence page: click a box, open the page behind it. Nothing leaves your site, and re-exporting after an edit does not mean redoing anything by hand.

And if your diagram is a simple drawing with no links and no small text, export it to PNG and move on. It is free, it takes a minute, and this whole article is beside the point for you.

Questions people actually ask

Is there a Visio macro in Confluence Cloud? No. There is no built-in macro that renders Visio files. The request has been open on Atlassian's tracker for years as CONFCLOUD-27877.

Can I at least show a thumbnail of the .vsdx? Not natively. The file downloads rather than previewing, so there is nothing to thumbnail. Some file-preview apps on the Marketplace add Visio support if you need the raw file viewable in place.

Does .vsd still preview? The old binary format behaved differently from .vsdx, which is the source of a lot of confusion in forum threads. Do not plan around it: saving modern diagrams back to a legacy format to win a preview is a bad trade.

What about multi-page Visio files? Export gives you one file per page, so a five-page drawing becomes five images and five uploads. The draw.io import handles multi-page .vsdx in one go, which is the main practical argument for importing when the drawing is large.

We are moving off Visio. Does that change the answer? It simplifies it: import once, keep the diagram editable inside Confluence, and stop exporting anything.

How to make an image clickable in Confluence Cloud

Mikhail Klimenko — Mon, 05 Oct 2026 20:27:34 +0000

There are two different questions hiding behind this one, and they have completely different answers. One takes two clicks. The other cannot be done with anything Confluence ships with, and the workaround most people reach for is the one that will cost them an afternoon in six months.

Short answer. To make the whole image a link: click the image in the editor, click the link icon in the floating toolbar, type a URL or search for a Confluence page, press Enter. Done. To make different parts of one image point at different places: Confluence Cloud cannot do it. That is an image map, image maps are built from HTML, and Cloud has no HTML macro. Your real choices are an app that draws hotspots over the picture, or a vector diagram that carries its own links.

Case one: the whole picture is one link

This is built in and takes seconds.

Click the image on the page in the editor. A small floating toolbar appears above it.
Click the link icon in that toolbar.
Type a URL, or start typing the name of a Confluence page and pick it from the dropdown.
Press Enter and publish.

Good for a logo that leads to a space, a screenshot that opens the tool it shows, a cover image that opens the document. If this is what you needed, you are finished.

Case two: different areas, different destinations

This is what people usually mean by an interactive picture: an architecture diagram where each box opens the runbook for that service, a floor plan where each room opens its own page, a process chart where every step links to its instruction.

On the open web this is an image map: an <img> tag with <map> and <area> elements marking out rectangles and polygons. In Confluence Data Center you could paste that into the HTML macro and be done. In Confluence Cloud the HTML macro does not exist, and raw markup is not rendered on a page. So the classic answer is simply unavailable, and no amount of searching will produce a setting that turns it on.

That leaves three real paths.

Slicing the picture into a table

The free workaround: cut the image into pieces in a graphics editor, upload each piece, put them in a borderless table, link each cell separately.

It works. It is also the option that quietly costs the most. Each slice is a separate attachment, so a change to the diagram means re-cutting and re-uploading every piece. Table cells introduce gaps and the picture stops looking like one picture. Column widths behave differently on a laptop and on a wide monitor, so the slices drift apart for some readers and overlap for others. And the next person to edit the page has no idea why the diagram is a table.

Reasonable for a picture that will never change again. A trap for anything that gets edited.

An app that draws hotspots

There are Marketplace apps built exactly for this, and they are the right answer for photographs and raster images. Scroll ImageMap from K15t gives you a graphical editor for drawing circles, rectangles and polygons over an image and pointing each at a page or URL. Smart Images does the same kind of job with rectangles, circles and polygons.

If your picture is a photo of a machine, a scanned floor plan or a rendered mock-up, this is what you want: there is nothing inside the file to work with, so somebody has to mark out the areas by hand, and these apps make that part pleasant.

A diagram that already knows where its links go

There is a case where drawing hotspots by hand is wasted work: when the picture is a diagram, and the links are already in it.

Draw.io, Visio and Mermaid all let you attach a link to a shape. Export that diagram as SVG and the links travel with the file, because SVG is markup rather than pixels. Then you upload it to Confluence and they stop working, because Confluence renders an uploaded SVG as a flat image. The information is in the file; the viewer just will not use it.

The annoying part is that this is the case where hand-drawn hotspots are pure duplication: you would be marking out, one by one, the areas the diagram already describes. And then doing it again after every edit, because the boxes have moved.

Disclosure: I build an app for this one. Clickable SVG renders the SVG as live markup instead of a flat picture, so the links your diagram already carries keep working: click a box, open the page behind it. Nothing leaves your site, and re-exporting an edited diagram does not mean redrawing anything. It does not help with photographs, where there are no shapes to read — that is what the hotspot apps above are for.

And if your case is the simple one, use what is already there. One link on the whole image is free, built in, and takes two clicks.

Which one fits your picture

Logo, screenshot, cover image, one destination — the built-in image link. Nothing to install.
Photo or scan, several areas, never changes — a hotspot app, or slicing if you are determined to spend nothing.
Diagram from draw.io, Visio or Mermaid with links on the shapes — keep the diagram vector and render it live, rather than redrawing its structure as hotspots.
Diagram that gets edited every month — anything hand-maintained will lose. Use the format that carries the links.

Questions people actually ask

Is there an HTML macro in Confluence Cloud? No. It exists in Data Center, not in Cloud, which is why image maps, custom scripts and pasted embed codes do not survive the move.

Can I link an image to an attachment or a file? Yes. The link control on a selected image accepts a Confluence page, an attachment or an external URL, the same as a text link.

My image map worked in Server. Why not now? Because it was rendered by the HTML macro, and that macro did not come to Cloud. The page will show the image, and the clickable areas simply will not be there.

Does a PNG exported from draw.io keep its links? No. PNG is pixels: whatever the shapes knew about their destinations is gone the moment you export. If you want the links to survive, export SVG.

Will the browser's own zoom break the hotspots? With hand-drawn hotspot coordinates, scaling is exactly where things go wrong, which is why the apps that do this properly store areas in relative terms. With a vector diagram the question does not arise: the shapes scale, and so do their links.

The Confluence Live Search macro: what it does, and where it breaks

Mikhail Klimenko — Mon, 05 Oct 2026 16:56:36 +0000

A search box sitting on a page looks like the obvious way to help people find things. It works, but not in the three ways most teams assume.

Short answer. Yes, Confluence Cloud has it. In the editor type /live search and insert the Live Search macro. You get a search field on the page, and you can restrict it to a space key, to one or more labels, and to a content type, plus set the field size, the placeholder text, and what appears under each result. What it does not do: it does not search the page it is sitting on, it cannot match the middle of a word, and it shows nothing at all until a reader types something.

What the macro actually is

The Live Search macro is a search field rendered inside page content. Someone starts typing, and Confluence suggests matching content underneath. The parameters Atlassian documents are these:

Restrict to this Space Key — a space key, case sensitive, or @self for the space the macro lives in.
Restrict to label(s) — limits results to content carrying those labels.
Size — medium or large search field.
Placeholder text — what the empty field says, for example Search this space.
Type — pages, live docs, blogs, comments, space descriptions, or all content.
Additional — whether to show the space name, a content excerpt, or nothing under each result.

That is a reasonable feature, and for a space landing page with a clear audience it does its job. The trouble starts where expectation and behaviour part ways.

Three things people expect and do not get

It does not search within the page

This is the most common misunderstanding, and it is baked into how people search for the feature: live search within page. The macro is placed on a page, but it searches the space, or whatever labels you restricted it to. The text of the page the reader is looking at is not what it looks through.

If the goal really is to find a word on the current page, nothing in Confluence beats the browser: Ctrl+F, or Command+F on a Mac. No macro needed, and it searches what is actually on screen.

It cannot match the middle of a word

The macro uses ordinary Confluence search, so it inherits a rule that surprises people. Atlassian's search syntax documentation puts it plainly:

Confluence doesn't allow wildcards at the beginning of your search. For example, you can't search for *hum* or ?hum*, as they begin with a wildcard.

In practice that means budget* finds budgeting and budgets, but there is no way to find annual-budget-2026 by typing budget if the word starts somewhere else in the title. The reader types a word they are sure is in the document, gets an empty dropdown, and concludes the page does not exist. It does. The search just cannot start in the middle.

This hits hardest in spaces where titles follow a convention, because conventions tend to put the useful word last: 2026-Q3-Finance-Budget.

It shows nothing until someone types

A search field is a question waiting for a question. It assumes the reader already knows what they are looking for and can name it in the same words the author used. New joiners, auditors and anyone arriving from a link do not meet that assumption. They look at an empty box and leave.

There is a second, quieter cost: a search box leaves no trace of what is in the space. A list does. Anyone glancing at a page with twelve linked documents learns the shape of the space in two seconds, without typing anything.

When a search box is the wrong tool

Use the macro when the reader knows the name of the thing and the space is large: a knowledge base with hundreds of articles and a clear naming scheme is a good home for it.

Reach for a list instead when any of this is true:

the reader does not know what exists and cannot name it;
the set of pages is defined by a rule rather than a word, such as everything in a space with a given label, or everything not updated in six months;
the page is meant to be an index that someone will link to and cite;
you want the page itself to show that the documentation is alive.

The hand-maintained index, and why it dies

The usual answer to all this is a page of links written by hand. It works beautifully for about a month.

Count the real cost. A team space gains and retires maybe eight pages a month, which is modest. Keeping an index honest means checking it roughly every two weeks: fifteen minutes to walk the space, find what is new, drop what has been archived, fix the titles that changed. Call it thirty minutes a month, six hours a year, for one index. Teams that take documentation seriously keep four or five of them.

What actually happens is that nobody does it. The index drifts for a quarter, somebody notices a dead link in a meeting, and it gets rebuilt from scratch. Then the cycle repeats.

A live list instead

The alternative is a list defined once by conditions and rebuilt by Confluence every time the page is opened: a space, a label, a content type, an age. Nothing to maintain, nothing to retype, and the page reads as an index rather than an empty box.

Disclosure: I build one of these. Our Smart Search macro puts that list on a page from plain conditions, without anyone writing a query language: choose a space, add labels, say not updated for 180 days, and the list keeps itself current. It is free for up to ten people, and it runs entirely inside your own site.

If your case is simpler than that, use what you already have. One space, a clear naming convention and readers who know the vocabulary — the Live Search macro is free, it is already in your editor, and it will do the job.

Questions people actually ask

Where is the Live Search macro in the new editor? Type /live search on a page and pick it from the list. It is a standard macro in Confluence Cloud, nothing to install.

Can I restrict it to more than one space? The space parameter takes a single space key, or @self for the current space. Covering several spaces means several macros, or restricting by label instead and labelling consistently across spaces.

Why do my results change depending on whether I press Enter or click through? With two or more labels this used to differ: pressing Enter required all labels to match, while the suggestion list showed pages with any of them. Atlassian tracked that as CONFCLOUD-53716, it collected 27 votes and is now closed as fixed. If you still see something like it, compare the suggestions against the full search results screen before assuming the pages are missing.

Can I make the search box appear in the sidebar? No. The macro lives in page content. The sidebar is built from the space navigation, and macros do not render there.

Does any of this change on Premium or Enterprise? Not for this macro. The search syntax rule about wildcards is the same on every plan, because it comes from the search engine rather than the subscription. What Premium adds is the space content manager, which is an admin screen rather than something you can put on a page.

Why can't someone see a Confluence page?

Mikhail Klimenko — Mon, 05 Oct 2026 02:20:20 +0000

You sent a link, they got an error, and you are now the support desk. The answer is almost always one of two layers, and checking them in the wrong order is what turns five minutes into an afternoon.

Short answer. Access is decided in two places: space permissions and page restrictions. Check the space first, because page restrictions can only take access away, never add it. Then check the page, and then check the pages above it, because view restrictions cascade downwards. The list of restricted pages in a space lives in space settings, under Content, in a section called Restricted.

The rule that settles most arguments

Space permissions win. Atlassian spells it out: "if someone lacks space-level permission to edit content, then that person won't be able to edit this content, even though this content's access is set to Open."

Read that twice if you manage a wiki, because it explains the most common failure in access requests. Somebody opens the page, sets access to Open, tells the colleague it is fixed, and the colleague still sees nothing. The page was never the problem. Page restrictions narrow what the space already allows; they cannot widen it.

So the order is not a matter of taste. Space first, page second. Doing it the other way round means you spend your time in the one place that cannot possibly be the cause.

The cascade, and the half of it that does not cascade

View restrictions flow downhill. In Atlassian's words: "If someone can't view a parent content item or folder, they won't be able to view any child content items under it."

This is why the page you are staring at can look completely unrestricted while the person still cannot open it. The restriction sits two levels up, on a parent nobody has thought about since last year, and everything underneath inherits the consequence.

Edit restrictions do not behave the same way: they do not cascade to children. That asymmetry catches people out in the opposite direction. A team locks editing on a section's front page, assumes the whole section is protected, and the pages below stay editable by everyone who has space access.

What this means in practice. When someone reports a page they cannot see, walk up the tree rather than around it. The page itself is the least likely place to find the cause, because if its own restrictions were the problem you would usually have noticed when you set them.

Finding the restrictions without opening every page

There is a list, and most admins never find it because it sits with the other space tools rather than with the page. In space settings, open Content and select Restricted. That gives you the restricted pages in the space in one screen.

It is the same corner of the interface that holds the list of hidden and orphaned items, which is worth knowing in itself: Content is where Confluence keeps the answers to "what is unusual about this space", and almost nobody opens it.

When you are the admin and still cannot get in

Space admins can see what exists in their space and change restrictions on it. That covers the ordinary case: somebody left, their page is locked, you need it back.

The bigger hammer is the admin key, available to organisation admins on Premium and Enterprise, which bypasses restrictions to manage access to any content. If you are on Free or Standard and the only person who could open a page has left the company, that hammer is not in the box, and the honest answer to your colleague is that this is a plan limitation rather than something you are doing wrong.

The five-minute diagnosis

In order, because each step rules out the one before it.

One. Does the person have view permission on the space? If not, stop here and fix that. Nothing on the page matters yet.

Two. Are they in the group you think they are in? Access is usually granted to groups, and "I added them" often means "I added them to a group that does not have this space".

Three. Is the page itself restricted? If viewing is limited to named people, everyone else is out, including colleagues who have full access to the space.

Four. Are any of its parents restricted? This is the step people skip, and it is the one that explains the cases that look impossible.

Five. Open space settings, Content, Restricted, and look at the whole picture. If the same few people appear on every restricted page, you do not have an access problem, you have a habit, and that is a conversation rather than a setting.

Frequently asked

Does sharing a page by link grant access? No. A link is an address, not a permission. If the two layers above do not allow it, the link produces exactly the error your colleague reported.

Why can they find the page in search but not open it? Search results and access are different things in several corners of Confluence, which is why a title can surface while the content stays shut. Treat seeing the title as no evidence at all about permissions.

We use restrictions a lot. Is that bad? It is bad when it is accidental. Restrictions set deliberately on a few sensitive pages are fine; restrictions inherited by a hundred pages because of one forgotten parent are how a wiki quietly becomes unusable.

Disclosure. I build small apps for Confluence, including Smart Search, which builds live lists of pages from plain conditions. It is worth saying plainly that none of them helps here: permissions and restrictions are not something an app on the Marketplace should be working around, and the built-in screens above do the job. This article exists because the question is common, not because I have something to sell you for it.

How to get a list of Confluence pages with owners and dates

Mikhail Klimenko — Sun, 04 Oct 2026 20:55:44 +0000

Somebody asks for "just a list of our pages, who owns them and when they were last touched". It sounds like a five-minute export. It is not, and knowing why saves you an afternoon of looking for a button that does not exist.

Short answer. Confluence Cloud has no single report that hands you pages, owners and dates in one file. Three routes get you close, and each gives up something: the content manager, which is the real answer but needs Premium; the space CSV export, which any space admin can run but whose contents Atlassian does not document; and search filters, which show you the list on screen without exporting it.

The content manager: the closest thing to the report you wanted

This is the tool to reach for first, with one condition. Atlassian is explicit: "The content manager is an admin tool available in Confluence Premium and Enterprise." On Free and Standard it is not there, and no amount of looking will find it.

What it filters by, in full: last active, pages with deactivated owners, statuses, labels, types, contributors, last updated as a date range, and created as a date range. Types covers more than pages: databases, folders, live docs, smart links and whiteboards.

The filter list is the useful part, because it answers the question behind the question. People rarely want an inventory for its own sake. They want pages whose owner has left the company, or pages nobody touched since a reorganisation, and both are one filter away.

Its bulk actions carry documented limits worth knowing before you promise a timeline: archive up to 500 items at a time, delete up to 100 at a time, change owner on up to 100 at a time. A thousand-page space is several passes, not one.

The space CSV export: available to everyone, undocumented inside

Any space admin can run it: "Space admins can export a space (or a group of individually selected pages in a space) to PDF, CSV, HTML, or XML." And the scope is generous: "For CSV, you will export everything you can view, including attachments and comments by default." A site admin exporting via CSV or XML gets everything, not just what they personally can see.

Here is the honest gap, and it is the reason this article exists. The documentation does not list what columns the CSV contains. It says the export is "best if you want to import a space into another Confluence Cloud instance", which is a statement about purpose, not about structure. Whether your file arrives with an owner column and a created date is something you find out by opening it.

So do that first, on a small test space, before you build a process on top of it. Ten minutes of checking beats a promise you cannot keep.

Search filters: fine for looking, useless for handing over

The third route is the one everyone tries first: search with filters, read the results, copy them somewhere. It works for a question you are answering yourself today. It falls apart the moment the answer has to be sent to somebody, kept for an audit, or looked at again next quarter, because nothing about it is saved.

That is the real split between these three routes, and it is not about features. A screen answers a question once. A file answers it to somebody else. A list that lives on a page answers it again next month without anyone rerunning anything.

The arithmetic, when the request is real

The usual reason this comes up is not curiosity. Somebody left, somebody is auditing, or a space is being handed to a new team.

Doing it by hand on a four-hundred-page space means opening pages to see who owns them, because the tree does not show it. At twenty seconds a page that is over two hours, and the result is stale the week after. On Premium the same job is a filter and a bulk action, measured in minutes. That gap, hours against minutes, is what the Premium plan is actually selling here, and it is worth naming when somebody asks why the upgrade exists.

On Free and Standard the honest plan is different: export the space to CSV once, look at what you actually got, and build the rest around that file rather than around what you hoped it contained.

Disclosure: I build Smart Search, which puts a live list of pages on an ordinary page, built from plain conditions such as a space, labels, or pages not updated for six months, and it works on any plan. It is a list that keeps itself current, not an export and not an inventory report, so if what you need is a file to send to an auditor, the CSV export above is the right tool and my app is not. If what you need is a page the team actually looks at, that is the case it was built for.

Frequently asked

Does the page tree show owners? No, and that is why this question keeps coming back. The tree shows titles and structure, so ownership and dates have to come from one of the three routes above.

What about the API? It is the right answer for a recurring report, and the wrong answer for a one-off request from somebody who needs it this afternoon. If this is going to be asked quarterly, write the script once. If it is asked once, do not.

Why does the content manager include whiteboards and databases? Because a modern space is not only pages, and a clean-up that ignores the other types leaves half the mess behind. Its Types filter is also the quickest way to see what a space is actually made of.

How to find orphaned pages in Confluence Cloud

Mikhail Klimenko — Thu, 24 Sep 2026 23:02:38 +0000

Short answer. Next to your space's name in the sidebar, select More actions (the three dots), then Space settings. Open the Content menu and select Hidden. That screen is the one Atlassian describes as the place where "Confluence lets you see a list of orphaned items in your space".

The menu item says Hidden. The thing you were looking for is called orphaned. That single mismatch is why this question has been asked on the Atlassian community for years by people who were, at the time, two clicks away from the answer.

Hidden and orphaned are not the same word for the same thing

They arrive at the same screen, and they mean different things, which is worth holding straight before you start cleaning anything up.

Orphaned is about links. Atlassian's definition is blunt: "If you do not have any incoming links, your page or live doc will be an orphan." Nothing points at it. It is not that the page is broken, it is that nothing in your documentation refers to it.

Hidden is about the tree. An item becomes hidden when it is moved to the root of the space, which, as the documentation puts it, "removes the item from your content tree so people can't find it by browsing through the tree". You do it deliberately: space settings, the Content menu, Reorder, then drag the item up to the same level as the home page.

So one is a consequence of how people wrote, and the other is a choice somebody made. The documentation treats them together and the interface shows them in one list, which is convenient right up to the moment you assume everything in that list got there the same way.

The dangerous half of the word "hidden"

It does not hide anything from anyone. Atlassian says it plainly: "Hidden (and orphaned) items still show up in search results, and can be accessed if someone has the page or doc URL."

If you moved a page out of the tree because it held salary bands or an unfinished policy, it is still in search and still opens for anyone with the link. The documentation offers the honest alternative in one line: "If you don't want anyone to find a page or doc, you don't need to hide it" — use restrictions, archive it, or delete it.

Why pages end up orphaned

Three ways, and only one of them is anyone's fault.

A page was created from a link that was later removed or rewritten, so the only thing that pointed at it is gone. A page was moved during a reorganisation and the pointers stayed behind. Or a page was never linked at all: somebody created it from the plus button, wrote it, published it, and moved on. That last one is the most common in practice, because creating a page in Confluence does not require you to say where it belongs.

The reason to care is not tidiness. An orphan is a page nobody arrives at by reading, which means it never gets corrected. It still comes up in search though, which is the worst combination: invisible to the people who maintain the documentation, visible to the person who searched for it today and believed it.

Where the clean-up stops

Finding the list is one click. Doing something with it at scale is where Confluence Cloud stops short, and it is better to know that before you promise a clean-up by Friday.

The bulk tool is the content manager, and Atlassian describes it as "an admin tool available in Confluence Premium and Enterprise". It is genuinely good at what it does: archive up to 500 items at a time, delete up to 100, change owner or status across a selection. Its documented filters are last active, pages with deactivated owners, statuses, labels, types, contributors, last updated and created dates.

Read that filter list again and notice what is not in it: orphaned, or hidden. The place that lists your orphans has no documented bulk actions, and the tool with bulk actions has no documented filter for orphans. Selecting every orphaned page in a space and clearing them in one pass is not a workflow the documentation describes, on any plan.

One more gap worth naming, because people expect it to exist: the Cloud documentation does not describe a screen that lists incoming links for a given page. Incoming links appear in it only as the definition of an orphan. If you remember that view from an older Confluence, check your own site before planning around it rather than trusting a memory or an article, including this one.

The arithmetic on a real space

A three-year-old space with a few hundred pages usually turns up somewhere between twenty and eighty items in that Hidden list. Opening each one, reading enough to judge it, and deciding between keep, link, archive and delete runs about two minutes an item once you are warmed up. Fifty items is a focused afternoon, and it is an afternoon that does not repeat, because the ones you keep get linked and stop being orphans.

What does repeat is the trickle: every month adds a few more, created from the plus button by people who were in a hurry. That is the part worth automating, and the honest version of automating it is a list somebody sees, not a report nobody opens.

What to do on Monday

Open the Hidden list in your busiest space and read the titles, not the pages. Titles alone sort most of it: obvious drafts, duplicates with a number on the end, and the two or three documents that turn out to be load-bearing and simply lost their link. Link those back into the tree first, because that is the change that stops somebody rewriting a policy that already exists.

Then check the same list against the calendar. A page that is both orphaned and untouched for a year is a safe archive. A page that is orphaned but edited last week is not junk, it is someone working outside the tree, and deleting it will cost you a conversation.

Disclosure: I build Smart Search, which builds live lists of pages from plain conditions, such as a space, some labels, or pages not updated for six months, and puts that list on an ordinary page on any plan. It is honest to say what it is not: it works from those conditions, not from link structure, so it is the tool for the stale half of this job, not for finding orphans. For orphans specifically, the built-in Hidden list is the right answer and it costs nothing. If that is all you needed, you are done.

Frequently asked

Do orphaned pages hurt our search results inside Confluence? They appear in search like anything else, which is the point of the problem rather than a side effect. The cost is not ranking, it is that outdated pages keep answering questions long after everyone stopped maintaining them.

Is archiving better than deleting? For anything with a trace of institutional memory, yes. Archiving removes an item from the sidebar and search results while keeping it, which is exactly what you want for last year's process that somebody will ask about in March.

Can I stop pages becoming orphans in the first place? Partly, and it is a habit rather than a setting: create pages from the parent page rather than from the plus button, so the tree places them for you. Nothing enforces it, which is why the Hidden list keeps refilling.

How to export a Confluence database to Excel

Mikhail Klimenko — Mon, 21 Sep 2026 20:06:38 +0000

Short version: you cannot export a Confluence database to Excel directly. The database exports to CSV, and Excel opens the CSV. That is a one-minute job, and everything worth knowing here is about what happens afterwards.

The steps

Open the database, open the more menu (the three dots) at its top right, choose Export, and pick a format. Atlassian documents three: CSV, HTML and PDF. CSV is the one for a spreadsheet, and Atlassian describes it as "ideal for importing data into other database tools or spreadsheets". There is no XLSX option anywhere in that menu.

One checkbox matters: Export with current view. Ticked, the export keeps the view you are looking at. Left alone, you get all the data in a plain table layout. If you spent an afternoon building a filtered view for the finance team, that tick is the difference between sending them their view and sending them the whole database.

What the database already does, before you export anything

People export to Excel for two different reasons, and only one of them needs Excel.

If the reason is arithmetic, look under the field first. A Confluence database can show calculations beneath a column: for number fields Sum, Average, Min, Max, Count values, Count unique values, Count empty and Percent empty; for other field types the four counting ones. You turn one on by picking Calculate under the field. A total that lives in the database stays right when the data changes, which a total in last month's spreadsheet does not.

What the documentation does not describe anywhere is formulas or calculated columns: there is no documented way to make one field compute from another. So if your case is a running cost column, the export to Excel is not a workaround, it is the only route.

The trap is on the way back

Exporting is easy. Bringing the data home is where people lose an evening, and Atlassian documents exactly why.

Import into a database takes CSV only. During that import, dates are read month first. The documentation is blunt about it: "Confluence databases assume month comes first for dates, so 12/06/2025 will be treated as December 6, 2025" rather than Jun 12, 2025. It also states plainly: "We do not support date field through autoconversion/autodetection." So a European-formatted date column comes back as a different set of dates, quietly, in the rows where the day is twelve or lower. The rows where the day is above twelve are the lucky ones, because those break loudly.

Two more documented surprises from the same page. Text columns can arrive as tag fields: "String fields in CSV may get converted to tag fields", which happens when most of the values repeat. And the first row of your CSV becomes the field names, with duplicates renamed automatically.

What to do. Before a round trip, switch date columns to an unambiguous form, ISO style, year first. Import one copy into an empty test database and look at February before you touch the real one. And when the import offers you a choice between creating new fields, adding to existing fields and replacing the entire database, read that dialog twice: the third option does what it says.

The arithmetic, if you do this monthly

One database is a minute of clicking. The cost appears when the answer lives in several of them.

Each database is exported from its own menu, one at a time, and each download lands as its own file. Twelve databases across a space is roughly ten to fifteen minutes of clicking and renaming, plus the merge afterwards, and it repeats every month because the export is a snapshot rather than a link. That is not a disaster, it is a standing appointment with an evening.

Worth knowing before you plan around it: databases are not available in the Atlassian Government environment, and the documentation lists other things they do not work with, including anonymous access, admin key, Atlassian Analytics, Atlassian Data Lake, data connectors, screen readers and keyboard navigation, and some Marketplace apps. Atlassian says that list is not exhaustive.

Two questions people expect answered here, and the honest answer to both is that the documentation is silent. It says nothing about which plans include databases: the only availability limit stated is the Government environment one. And it does not say whether databases are included in a space export, in either direction; the space export page lists blog posts, comments and Team Calendars as exclusions and never mentions databases. If either point matters to your migration, test it on one space rather than trusting an article, including this one.

If your data is not in a database at all

Half the people who search for this do not have a database. They have ordinary tables on ordinary pages, which look similar and behave nothing alike: no export menu, no calculations under the column, and no CSV.

Disclosure: I build Table Export, which covers that second case, taking one table, a page, or every table in a space into a single workbook with a sheet per table. It does not replace the database export: if your data is in a database, the built-in CSV export above already does the job and you do not need my app. That is the honest split, and it costs me nothing to say, because half of these searches genuinely end at the Atlassian menu.

Frequently asked

Does the export include the calculations shown under a field? The documentation does not say. Treat the export as the rows, recreate totals in the spreadsheet, and check the first file you send anyone.

Can I get the data out without downloading files by hand? Not through the documented database export, which is a per-database menu action. If you need it scheduled, the honest answer is that this is a script-and-API job, not a menu one.

Why not just use HTML export? It is genuinely useful for one thing the CSV cannot do: Atlassian notes the HTML export gives you the option to copy the data to the clipboard, which is faster when you only need to paste a small table into an email.

How to make a diagram clickable in Confluence Cloud

Mikhail Klimenko — Sat, 19 Sep 2026 14:52:21 +0000

Short answer. If the diagram lives in a diagramming app such as draw.io or Gliffy, select a shape and add a link to it: both support links on shapes, pointing at a Confluence page, an external URL or another diagram. If the diagram is a plain image, Confluence can make the whole image one link, but not different parts of it. If the diagram is an SVG file you attached, the links inside it will not work on the page.

People have been asking this on the Atlassian community for more than a decade, and the answer splits cleanly in two. It is a two-minute job if the diagram was drawn inside Confluence, and a different problem entirely if the diagram arrived as a file.

draw.io: select the shape, Edit Link

Open the diagram in the editor, click the shape, then right-click and choose Edit Link. The same command sits in the Arrange tab of the format panel on the right, and the keyboard shortcut is Alt+Shift+L.

The dialog offers more than a plain URL. A web link opens in a new browser tab. Page Content lists the headings of the Confluence page the diagram sits on, so a box can jump the reader to a section further down. That list is generated from the published page, so if it comes up empty, publish the page first and reopen the dialog.

Readers get a hint for free: hovering a linked shape outlines it, so people discover the diagram is navigable without being told.

Gliffy: link the shape or the text inside it

Gliffy does the same job from its link toolbar, on a shape or on a piece of text, pointing at a web address, a Confluence page, an attachment or another Gliffy diagram.

One detail from their documentation is worth copying as a house rule, because it prevents a class of dead links. When linking to a Confluence page, use the Search option in the Insert Link window rather than pasting the address from the browser. Gliffy then stores the short Atlassian link, the ugly /x/ZygVQ kind, which keeps pointing at the same page after it is renamed or moved to another space. A pasted URL contains the page title and breaks the day somebody rewrites it.

A plain image gets one link, not several

If what you have is a PNG of a diagram, select it in the editor and use the link icon in the floating toolbar. That turns the picture into a link, the whole picture, as one target. It is the right tool for a banner or an icon that should open one place.

What it cannot do is give each box its own destination. Confluence Cloud has no hotspot editor for regions of an image, so an architecture map saved as a PNG stays a picture of a map.

Why an attached SVG loses its links

This is the case that sends people looking for an answer, because an SVG genuinely does contain links. You exported it from Figma, Visio, Inkscape, Graphviz or an architecture tool, the file opens in a browser and every box works. Attach it to a Confluence page and it becomes a flat picture.

The reason is how the file is displayed. Confluence Cloud shows an attached SVG through its media viewer as a rendered picture, not as live markup inside the page. The separate suggestion to allow SVG tags in a page storage format, CONFCLOUD-36336, "Allow SVG in Storage Format", has been open since January 2015 and is still marked Gathering Interest.

The confusing part, worth knowing before you plan around it

The ticket most people find is CONFCLOUD-1762, "Embedding of SVG", opened in 2004 and carrying 367 votes, and it was closed as Fixed on 10 September 2024. That looks like the answer, and it is not. What shipped was the new SVG rendering engine announced in October 2023, which addressed SVG 1.2 support, text dropping out and poor rendering of Mermaid diagrams. Better viewing, not interactivity. The product manager's note at the time also asked people to reattach existing SVG files so the new engine would redraw them, which is good advice if your diagram looks wrong and beside the point if your problem is that clicking does nothing.

The arithmetic, if you are deciding whether to redraw

There is a free way out of the SVG case. draw.io imports SVG files, along with VSDX from Visio, Lucidchart files and PNG, through File → Import from → Device or by dragging the file onto the canvas. Import the diagram, add the links inside draw.io, and you have a clickable diagram with no extra app.

Work out whether that is a one-off or a habit. Adding a link to a shape takes twenty to thirty seconds once you know where the dialog is, so a forty-box service map costs fifteen to twenty minutes. Paid once, that is nothing. The question is what happens on the next export.

If the diagram is drawn by a human and changes twice a year, import it and forget this article. If the SVG is generated, by Graphviz, Structurizr, a Mermaid or PlantUML build step, or any pipeline that redraws the picture from the system it describes, then every regeneration produces a file with the links already in it and no links in your Confluence copy. You are back to fifteen minutes of clicking, on a schedule. That is the case where redrawing is the wrong answer, because you are hand-maintaining something a machine already produced correctly.

What to do about the exports

Whatever route you take, do not let clicking a box be the only way to reach the page behind it. Diagrams come out of a PDF export as pictures, and a printed runbook whose navigation lived in the diagram is a runbook with no navigation. Put the important destinations in a list under the picture as well. It costs one line, and it also makes those links visible to search, which pixels are not.

Disclosure: I build Clickable SVG, an app for exactly the last case, an SVG that already has its links and needs them working on a Confluence page without being redrawn. If your diagram does not exist yet, or exists once and never changes, draw it in draw.io or Gliffy and link the shapes there. That is free, it is the path the Atlassian ecosystem is built around, and you will not need my app. If your case is simpler, take the simpler option.

Frequently asked

Can a shape link to a heading on a different page? Link to the page and add the anchor to the address yourself. The heading picker in the editor works on the page the diagram sits on, which covers the common case of a diagram acting as a table of contents for a long document.

Is there a way to keep one diagram and use it on several pages? Both diagramming apps support reusing a diagram across pages rather than copying it, which matters more than it sounds. A copied diagram is a diagram that will be updated in one place and stale in three.

What about diagrams generated from code? Text-based diagrams are the strongest option when the picture has to track a system that keeps changing, since the build redraws it. Decide in advance how the links get in, because that is the part the generator usually does and the page usually loses.

Does any of this depend on the Confluence plan? The behaviour described here is about how diagrams and attachments are displayed, not about plan tiers. App pricing is a separate question and changes over time, so check the current listing of whichever diagramming app you are considering.

Confluence databases or a table on a page

Mikhail Klimenko — Wed, 16 Sep 2026 18:13:06 +0000

Short answer. Use a table when the rows belong to the text around them: a budget inside a proposal, a comparison inside a specification. Use a database when the rows are the point and people will filter, sort and keep adding to them: a register of documents, assets, risks, suppliers. And before you commit either way, check how the data comes back out, because that is where both of them disappoint people.

What a database actually gives you

A Confluence database is a separate piece of content that lives in the space tree, next to pages, with permissions handled at the space level like everything else. You create it from the Create menu, from the plus next to Content in the sidebar, or by typing /database.

What you get that a table cannot do:

Typed fields. A date is a date, an owner is a person, a link is a link. Confluence knows what the value means instead of treating everything as text.
Filters. Filter by a field with real conditions, is, is not, contains, and work with the result.
Saved views. The same data shown several ways, each view keeping its own filters and sorting, so "open items only" and "everything by owner" are two clicks, not two tables.
Aggregates. A sum or an average at the bottom of a field, saved as part of the view.
One source, many pages. The database can be shown on the pages that need it instead of being copied into each of them.

What a plain table gives you

Less, and that is sometimes the right answer. A table sits inside the page, in the flow of the text, and is written and read together with it. It takes any layout you like, including merged cells and mixed content in a cell. Readers can sort it by a column while viewing the page, ascending or descending.

No setup, no separate object to find, no permissions of its own to think about. For a table that exists to support the paragraph above it, that is the whole requirement.

Where the database stops

Limit	What it means in practice
No per-row formulas	You cannot make a column that computes amount multiplied by rate. Aggregates at the bottom of a field exist, per-row calculation does not. Atlassian has an open suggestion, CONFCLOUD-78459, and until it ships the workaround is to calculate elsewhere and type the result in.
Export is manual	CSV, HTML or PDF, exported by a person from the database. Fine occasionally. There is no practical way to pull the content through the Cloud REST API, so a scheduled export or a feed into a reporting tool is not something you can build today.
Size	No published row limit, but large databases get slow, and community reports describe imports of a few thousand rows with many columns failing outright. Test with your own data before migrating a big register.
It is not a spreadsheet	People arrive expecting Excel behaviour, cells referring to other cells, and that model does not exist here at all.

The formula gap is the one that catches teams out. A register of invoices where the total has to be quantity times price looks like an obvious fit for a database right until you try to build that column.

Where the table stops

No filters. Sorting by a column exists for readers, filtering does not. A hundred-row table is a hundred rows for everybody.
Merged cells switch sorting off for the whole table, and merged cells are exactly what people use to make a table readable.
Nothing is typed. A date in a cell is a piece of text that looks like a date, which is why the same column ends up holding 16/09/2026, Sep 16 and "next Tuesday".
No export at all. Confluence exports whole pages to PDF and Word, and has no built-in way to send a table to Excel or CSV.

How to choose, in one question

Ask what happens if you delete the surrounding page. If the rows should survive, they are a register and belong in a database. If the rows make no sense without the paragraphs around them, they belong in a table on that page.

A second question worth asking early: who needs this data outside Confluence, and how often. Monthly, by hand, from one place is a different answer from every week, from forty pages, into a finance spreadsheet.

Getting the data back out

This is the part that decides more cases than any feature comparison. A database exports itself to CSV, HTML or PDF, which covers the occasional manual export and nothing beyond it. A table on a page has no export at all: you select it, copy it, paste it into a spreadsheet, and discover that merged cells shifted the columns after the first merge, silently, while the numbers still look perfectly reasonable.

For one table today, copy and paste and be done. The arithmetic changes when it is the same tables every month, from pages spread across a space, and when the shifted column ends up in a report that somebody signs.

Disclosure: I build Table Export for Confluence, which covers that last case: one table, a whole page, or every table in a space into a single Excel workbook, each sheet labelled with the page it came from, with merged cells expanded so the columns stay aligned. It works on tables in pages, not on databases, which export themselves. It runs inside Atlassian's infrastructure, so page content is read in your own site and nothing is sent to me.

If your data already lives in a database and you export it twice a year, use the built-in CSV and spend nothing. Most cases genuinely are that simple.

Questions people ask

Can a database replace a spreadsheet?

Not while there are no per-row formulas. It replaces a list you were keeping in a spreadsheet for lack of anywhere better, which is a large share of such spreadsheets, but it does not replace a model where cells depend on each other.

Can I show a database on a page?

Yes, and that is the normal way to use one: the data lives in the database, and the pages that need it show it. That is the main structural advantage over tables, which have to be copied.

Do database permissions work separately from the space?

Database access follows Confluence content permissions and is managed at the space level, the same as pages. There is no separate permission system to learn, which also means no per-row visibility.

Should I migrate my existing tables into databases?

Only the ones that are registers. Moving a table that explains a document into a database makes the document worse and helps nobody. And test with a copy first if the table is large.

How to move a Google Doc into Confluence Cloud

Mikhail Klimenko — Tue, 15 Sep 2026 21:22:47 +0000

Short answer. Confluence Cloud imports Google Docs directly. Start creating a page, open More actions next to the Share button, choose Templates and import, open the Import tab, pick Google Doc and authorise your account. The document arrives as a Confluence page. For a short document that really is the end of it. Shared drives, images and any document long enough to deserve more than one page are where the trouble lives.

The path that works

Select Create in the top navigation, or the plus next to Content in the sidebar. This matters more than it sounds: the import belongs to page creation. If you are sitting in an existing page looking for an import button, you will not find one, and nothing on screen explains why.
Next to Share, select More actions, then Templates and import.
Open the Import tab and choose what you are bringing in: a Word document, a Google Doc or a OneDrive document.
Authorise Google and pick the file.
Select Finish. Each file you import becomes its own Confluence page.

Several documents can come in at once. Atlassian is explicit that the requirements differ between importing one file and importing a batch, and that they vary by document type, so the sensible move is to test with your own files rather than plan a migration around a number you read somewhere.

What the built-in import gets right

It deserves credit. There is no download, convert and re-upload dance: Confluence talks to Google directly, and the file never lands on your laptop. Body text, lists and most tables come through in a state you can work with. Bringing in a folder of separate documents, each becoming its own page, is genuinely quick. If your job is "we have forty meeting notes in Drive and they should live in the wiki", this feature does that job and you can stop reading here.

Where it breaks

Situation	What actually happens
Document on a shared drive	The picker shows your personal Drive and the document simply is not in the list. This comes up repeatedly in the Atlassian Community and there is no setting on the Confluence side that fixes it.
"This file can't be imported"	A flat refusal with a generic explanation about file type or permissions. Usually the wrong type, or a document you opened through a link without real access to it.
Images	Usually fine, sometimes not: community reports describe pictures arriving as placeholders where the image should be. Always look at the imported page before you delete anything.
Heading levels	The document title becomes the page title, and everything under it shifts up a level. Your Heading 2 sections arrive looking like top-level headings.
Shapes and drawings	Atlassian documents plainly that shapes in Word documents are replaced with placeholders. Diagrams drawn inside the document do not survive as diagrams.
Anything long	One file becomes one page, however long the file was.

The failure that costs money is not the obvious one. A refused import is visible and you deal with it. A page that imported cleanly except for four missing images looks finished, gets linked from three other pages, and the gap is found by the person who needed exactly that diagram.

The .docx detour

When the Google path refuses, go around it. In Google Docs choose File, Download, Microsoft Word (.docx), then import that file through the same Import tab as a Word document. Only the .docx extension is accepted.

This is the standard answer in the community for shared drive documents, and it is worth knowing for a second reason: it gives you a file you can edit before importing. If the document needs cutting up, and long ones do, you have to be holding a file to cut it.

One file, one page, and why that hurts

Confluence Cloud creates a single page per imported document. That is fine for meeting notes and wrong for anything structured. A sixty page handbook arrives as a single page of endless scroll, with no navigation, nothing anyone can link to by section, and a table of contents that is just a list of anchors on one giant page.

The option to split a document into several pages by heading level does exist in Atlassian's own product line, but in Confluence Data Center, where the import screen offers one page or multiple pages built from heading levels. Cloud does not have that screen. People arrive from Data Center, go looking for the choice, and conclude they have missed a setting. They have not.

How long the manual way takes

Say the handbook has twenty five sections that each deserve a page. Per section: create the page, copy the section across, fix the headings that shifted, put back the images, place it in the right spot in the tree. Six to eight minutes when nothing fights back. That is roughly three hours for the document, and the part that stings is that it comes back on the next revision, because the source of truth is still the document and somebody will rewrite four sections in it next quarter.

What to do with the long ones

Two honest options. Cut the document into separate files in Word or Docs first, one per intended page, and import them as a batch: tedious but free, and completely predictable. Or use an app that does the splitting during import, which is worth it only when this happens more than once.

Disclosure: I build Word Import for Confluence, which reads the headings in a .docx and creates the page tree from them in one pass, instead of one long page. It runs inside Atlassian's own infrastructure, so the file is processed in your site and nothing is sent to me. It is an app for the structured case, and it is currently going through Marketplace review.

If your case is a folder of meeting notes, use the built-in import and spend nothing. The free path is genuinely good at the job it was built for, and most documents never need more than it.

Questions people ask

Where is the import if I am already editing a page?

It is not there. Document import happens while creating a page, through More actions, Templates and import. Editing an existing page gives you no import option, which is the single most common reason people conclude Cloud cannot import at all.

Can I import straight from a Google shared drive?

Often not. The picker tends to show the personal Drive only. Download the document as .docx and import it as a Word file.

Does the imported page stay connected to the Google Doc?

No. The import is a one-time conversion. Edits made afterwards in Google do not reach Confluence, and edits in Confluence do not reach Google. If you need the document to stay live in its original home, embed it with the Google Drive macro instead of importing it.

How many documents can I import at once?

More than one, and the requirements differ between a single import and a bulk import and depend on the document type. Atlassian says so without publishing one number that covers every case, so test on your own files before committing a migration plan to a figure.

Can I import a PDF?

Not as page content. The import handles Word, Google Docs and OneDrive documents. A PDF can be attached to a page, which is a different thing: an attachment is a file sitting on a page, not text anyone can edit or search inside the way page content is searched.

The Confluence content manager: what it does and where it stops

Mikhail Klimenko — Sat, 12 Sep 2026 03:20:10 +0000

Short answer. The content manager is a screen inside a Confluence space's settings. It lists everything in that space except blogs, filters the list by last activity, labels, status, type and contributor, and then archives, deletes, re-owns or re-statuses hundreds of items in one action. It is available only on Confluence Premium and Enterprise: if it is missing from your Space settings, that is almost always why. Its Last active filter is the fastest native way to find forgotten pages, and it carries one catch the screen never mentions: it leaves out every page that has children.

Where it is

Open the space. Next to the space name in the sidebar choose More actions, the three dots, then Space settings, then Content manager. You need to be a space admin of that space. It works on one space at a time, and Atlassian is specific about the scope: a space admin can use it to organise all of a space's content except blogs.

That last detail matters more than it sounds. If your team publishes announcements and decisions as blog posts, the content manager will not show you a single one of them, no matter how old.

What it can filter

Filter	What you can ask for
Last active	Four fixed periods: over 7 days, over 1 month, over 3 months, over 6 months
Last updated	A date range you supply
Created	A date range you supply
Labels	Content carrying particular labels
Statuses	Content with a given page status
Types	Databases, folders, pages, live docs, smart links, whiteboards
Contributors	Content a given person worked on
Pages with deactivated owners	Content owned by people who have left

Notice the shape of that list. Last active is the only filter that thinks in relative time, and it does so in four fixed steps. Last updated and Created want real dates, which is a different question: "updated between March and June" is an investigation, "not touched for six months" is an audit.

The filter worth knowing precisely is Last active. It means no views, no updates and no comments in the period. That is a better definition of stale than "last edited", because a page nobody edits but everybody reads is not dead, it is finished.

What it can do to what it finds

This is the half that no macro and no app replaces, and it is genuinely good:

Archive up to 500 items at a time. Archived content stays viewable but stops cluttering the content tree.
Delete up to 100 items at a time, into the trash, where an admin can purge it permanently.
Change owner of up to 100 items at a time. This is the fastest fix after somebody leaves the company.
Change status, run an automation flow against the selection, and convert pages to live docs.

Permissions survive the trip. Atlassian states that after a nested group is archived, inherited permissions remain intact in the archive, and again in the content tree if it is restored. Selecting a parent's checkbox automatically selects everything nested under it; hold Command or Alt to select a parent without its children, and Shift to take a run of rows at once.

The catch: your stale list is shorter than the truth

Atlassian puts this in a warning box, and it is easy to read past: Last active will not return a complete list, because it specifically filters pages that do not have children.

The reasoning is sound. Nested content inherits view permissions from its parent. If a parent or a child is moved without permissions being set explicitly first, content can be exposed to an audience it was never meant for. Excluding parents from a bulk-action screen is a reasonable way to avoid that.

The effect on an audit is another matter. Think about which pages actually rot in a real wiki. It is rarely the leaf: it is the section landing page. "Onboarding", with six pages under it, last touched when the process changed two reorganisations ago. "Release process", with a child per environment. Those are all parents, and none of them will appear in your Last active results. You will get the leaves, act on them, and come away believing the space is in better shape than it is.

There is a way to act on a whole nested group once you have found it by other means: open the parent's More actions menu and choose to archive or delete, and you are offered the option to include nested content. Finding them is the part you are on your own for.

If you are on Free or Standard

Then you do not have this tool at all. Atlassian's wording is blunt: teammates and administrators of Confluence Standard do not have access to it. What Standard does give you is the ability to archive or delete a nested group directly from the content tree, one group at a time.

That leaves three honest options. Sort a search by last modified and work through it by eye. Use the filter by label macro, which accepts twelve filters from a form and is stronger than its name suggests, but whose date filters point at a fixed date rather than a relative period. Or use an app.

What it is worth in time

Take one space with 1,500 pages. Working through a date-sorted list by eye, at three seconds a row to decide whether a page still matters, is about 75 minutes. Do that quarterly and it is five hours a year, for one space. Twelve spaces and it is a job nobody will do twice.

The content manager collapses the finding part of that to one dropdown, and the acting part to one click on 500 items. That is why it is worth knowing whether your plan includes it before you budget a documentation clean-up as a manual project.

The honest summary

If you are on Premium or Enterprise and you want to clean up one space, open the content manager and use it. It is native, it is already paid for, its definition of stale is the right one, and its bulk actions are better than anything a macro can do. Just remember that the list in front of you has no parents in it.

Where it stops is equally clear. It is one space at a time. It is an administrator's screen, so the people who wrote those pages never see the list and never get the nudge. It is not on Free or Standard. And it cannot be put on a page.

Disclosure: I build Confluence apps. One of ours, Smart Search, exists for the gap in that last paragraph: a relative period like "not updated in six months", on an ordinary page any reader can open, across several spaces at once, on any plan. What it deliberately does not do is bulk archiving or deleting. If you have Premium and you want a one-off clean-up of a single space, the content manager is the better tool for that job, and you are already paying for it.

Questions people actually ask

Why can I not see Content manager in my space settings? Because it is a paid feature, available on Confluence Premium and Enterprise only. It is not hidden or misplaced. You also need to be a space admin of the space you are looking at.

How far back does Last active go? Four settings: over 7 days, over 1 month, over 3 months, over 6 months. Six months is the limit, and you cannot type a number of your own.

Why are old pages missing from my results? Last active excludes any page that has children, as a safeguard around inherited permissions. Parents never appear, which is why the list looks better than reality.

Does it cover blog posts? No. All of a space's content except blogs.

How many items can I act on at once? Archive 500, delete 100, change owner 100.

Does archiving break inherited permissions? No. They stay intact in the archive, and again if the content is restored.

Can I show this list on a normal page? Not with the content manager. It is an admin screen, not a macro.

Can you import a Word document into an existing Confluence page?

Mikhail Klimenko — Sat, 12 Sep 2026 03:18:24 +0000

Short answer. In Confluence Cloud, no. The import always creates a new page, because it is an entry point to page creation rather than something you do to a page that already exists. Confluence Data Center can import a Word document into an existing page; Cloud cannot. Your practical options are to import as a new page and move the content across, or to paste the content into the existing page, which is uglier but keeps the page's history, comments and URL.

Why the button is not where you expect

In Cloud you reach the import while creating a page, not while editing one. A single file becomes a draft or live doc that you then publish. Several files at once land as child items under a restricted placeholder page titled "Imported pages" with the date and time. Either way the result is new pages, never an update to an old one.

This surprises people mostly in one situation: a document that lives in Word, gets revised once or twice a year, and has a Confluence page that everybody links to. You want the new revision on the same page. Confluence wants to give you a new page.

The four ways round it

1. Paste into the existing page

The only route that genuinely keeps the page. Open the page, select everything, paste the new content, publish. You get a new version of the same page: the URL still works, the comments are still attached, the history shows what changed and when.

The cost is the paste itself, which is covered below and is not free.

2. Import as a new page and move the content

Import the document, then copy the rendered content from the new page into the old one and delete the new page. It sounds like a detour, and it is, but the intermediate page is Confluence's own conversion of the document, which is usually cleaner than what the clipboard produces.

3. Import as a new page and retire the old one

Fastest, and the one people regret. The new page has a new URL, no comments and no history. Every link in other pages, in Jira tickets and in old emails now points at the previous revision or at nothing. Only reasonable when the page is genuinely new work that happens to have an old namesake.

4. An app

Several Marketplace apps extend importing in various directions. Worth pricing against how often you actually do this: for a document revised twice a year, a licence is hard to justify.

What pasting from Word actually does

Word does not put plain text on the clipboard. It puts markup, and Confluence keeps it: span and div elements carrying styles that the editor then has to live with. The visible result is usually fine at first glance and wrong in the details.

Lists are the usual casualty. Numbered and bulleted lists coming from Word get their numbering mixed up on paste; it is one of the more common complaints in the community. Nested lists fare worst.

Plain text paste is clean and expensive. Pasting without formatting gives you a page with no hidden junk and no headings, no bold, no lists and no tables either. For a short document that is a fine trade. For a fifty page handbook it is an afternoon of re-formatting.

Markdown is the useful middle. Confluence Cloud auto-converts pasted Markdown, and headings, bold and italic, ordered and unordered lists, links, images from URLs and simple tables come across reliably. Merged cells and fussy alignment still need the table editor afterwards. If your document can go through a Word to Markdown conversion first, this is usually the cleanest paste available.

What it costs on a real revision cycle

Say forty policy documents, each revised once a year, each with a Confluence page people link to. Pasting and cleaning up one revised document takes fifteen to twenty minutes when lists and tables are involved. That is ten to thirteen hours a year, spread thin enough that nobody counts it, and concentrated enough in one week of the year that it always lands on one person.

The honest observation is that no tool removes this entirely, because the problem is not the import, it is keeping the same content in two systems.

Which route to take

If the page matters to other people, paste into it and keep the history. A slightly messy page with its comments and links intact beats a clean page nobody can find.

If the document is long and full of structure, go through Markdown rather than pasting from Word directly.

If this happens once, stop optimising and just do it by hand.

Disclosure: I build Confluence apps, and this is a case ours does not cover. Word Import turns .docx files into new Confluence pages and page trees; it does not update a page that already exists, so for the question at the top of this article it will not help you. If what you actually need is to bring a pile of documents in for the first time, that is the case it is built for. If you need to refresh one existing page, use option one above and keep your history.

Questions people ask

Can I import a Word document into an existing page?
Not in Cloud. The import always creates a new page. Data Center can import into an existing page.

Why does Cloud work this way?
The import is part of page creation, not editing. One file becomes a draft you publish; several land under a restricted "Imported pages" placeholder.

What breaks when I paste from Word?
Hidden markup comes along, and list numbering is the usual victim. Plain text paste avoids it and loses all formatting.

Is there a cleaner way to paste?
Convert to Markdown first. Confluence auto-converts it, and headings, lists, links and simple tables survive.

Does the page keep its history if I replace the content?
Yes, if you edit the existing page. It becomes a new version, and the URL, comments and history stay. Importing as a new page and deleting the old one loses all three.