« Payment isn’t going through. » The support ticket says nothing more. On the backend, no error spike; on the frontend, a few JavaScript exceptions among thousands of others. Without seeing the customer’s screen, the team is left guessing, and the ticket risks ending up as « cannot reproduce ». This is exactly the case Dynatrace Session Replay was rebuilt for.
Session recording tools have met this need for a long time, but they usually live on their own: you watch the video in one tool, look for the trace in another, and copy a request ID into a third. The new Dynatrace Session Replay, now available to all SaaS customers on Grail, changes the game: every recording is stored alongside the logs, traces, events and RUM data of the same platform.
This article looks at what this brings in practice, how to get started, and what to watch out for if you still use Session Replay Classic.
Dynatrace Session Replay: from film to investigation
A replay on its own shows what the user did. It does not say why the button did not respond. The value of the new version lies in its foundation on Grail: a session is no longer a video file, it is a set of queryable events linked to the rest of the telemetry.
In practice, this opens several investigation paths, in both directions:
- from a session event to the matching distributed trace, span by span;
- from a log line to the user sessions concerned;
- from a backend incident to the sessions active at the same time;
- from any DQL query to Notebooks, Dashboards, Workflows or Investigations.
No more copying request IDs from one tool to another: the frontend-backend link is built in.
What the player shows
During playback, the chronological list of events scrolls next to the recording: requests, errors, user actions, navigation. A timeline groups these events and lets you jump straight to a specific moment; a visual marker flags the key moments. The usual controls are there: playback speed, skipping inactive periods, going back a few seconds. Each event expands to show its detailed attributes, including the trace ID, and a search bar lets you navigate long sessions without watching them in full.
A summary by Dynatrace Intelligence
The Summarize Session button asks Dynatrace Intelligence to read the session’s RUM events and produce a plain-language summary: what the user experienced, the anomalies and errors detected, and what happened beneath the surface. For a twenty-minute session, it is often the quickest way to know whether it is worth watching.
Beyond the incident
Extended retention of Session Replay data is available in preview. It opens up uses beyond immediate troubleshooting: handling customer complaints received weeks after the events, compliance audits, fraud detection. Because digital experience data stays in Grail, you can also analyse trends across session history with DQL, as you would for logs with log pattern analysis.
Another often underrated use: non-technical teams. Customer support and business teams readily use a replay to understand a complaint, provided access is properly scoped.
In practice
Step 1 — Enable recording, ideally as configuration as code
For a few applications, enabling it manually is enough. Beyond that, use Dynatrace configuration-as-code tooling to apply the same recording policy to all your frontend applications, without tuning each application individually. This is also the right time to review masking rules for user input: a replay must never expose a card number or sensitive personal data.
Step 2 — Find the right sessions
Open the Users & Sessions app and apply the filter for sessions with a replay, which are clearly flagged in the list. Facets let you narrow down by user ID, country, browser, operating system, frontends visited or number of errors. For a support ticket, the user ID and the time range are usually enough.
Step 3 — Go further with DQL
When facets are not enough, query user events directly. Example: count the sessions of real users who reached a checkout page in the last 24 hours, excluding bots and synthetic tests.
fetch user.events, from: now()-24h
| search "checkout"
| filter dt.rum.user_type == "real_user"
| summarize sessions = countDistinct(dt.rum.session.id)
By replacing the last line with dedup dt.rum.session.id, you get the list of distinct sessions, which you can open one by one. These results appear directly in a Notebook (see the training module on Notebooks and Dashboards), which keeps the analysis and its conclusions in the same place. The semantic dictionary in the documentation lists the available fields.
Step 4 — Pivot to the backend
In the replay, find the failing action, expand the event, then follow the link to the distributed trace. You will see whether the request reached the backend, which service responded and with which error. If the request does not appear, the problem is most likely on the client side: that is already a diagnosis. For this link to work with services instrumented by OpenTelemetry, the chain must be complete: see end-to-end OpenTelemetry.
Step 5 — Migrate from Session Replay Classic
If you use Session Replay Classic, your existing configuration stays active and data is not billed twice for double storage. However, data access permissions must be granted for the new experience. Run both in parallel during the transition, and use the opportunity to define who, between support and technical teams, should see what.
Best practices
- Validate masking rules with your DPO before extending recording, especially if extended retention is being considered.
- Grant permissions in a targeted way: support needs replays, not necessarily application logs.
- Use the session summary to triage, then watch the recording to confirm.
- Always filter on
dt.rum.user_typeto keep bots and synthetic tests out of your analyses. - User Interactions, which represent each interaction as a standalone event, are in Early Access and must be enabled separately: test them on a pilot application.
Frequently asked questions
Does Session Replay record what users type?
Masking rules decide what is recorded. Configure them before extending recording, so that no card data or sensitive personal data ever appears in a replay.
Should we drop Session Replay Classic straight away?
No. The existing configuration stays active and both can run side by side during the transition; you only need to grant data access permissions for the new experience.
How do I find the session of a customer who opened a ticket?
In the Users & Sessions app, filter on sessions with a replay, then on the user ID and the ticket’s time range. For finer criteria, query user events with DQL.
Conclusion
The new Session Replay turns session recording from visual evidence into the entry point of a full investigation. Because it is built on Grail, a session naturally links to the trace, log or host that explains the observed behaviour, and vice versa. For teams that juggled a replay tool and an observability platform, it changes daily work; for everyone else, it is a good reason to finally look at what their users see.
Further reading: Investigate user sessions end-to-end with the new Session Replay (Dynatrace blog).
span.observer is an independent community blog, not affiliated with Dynatrace LLC. Dynatrace is a trademark of Dynatrace LLC. SaaS features evolve quickly: check their availability in your environment and in the official documentation.
Originally published on span.observer. French version: https://span.observer/fr/session-replay-grail/.
Top comments (0)