Most multi-cloud database discussions start and end with AWS, GCP, and maybe MongoDB Atlas. On-premise databases rarely come up — not because teams don't have them, but because almost no tooling treats them as a first-class citizen.
That's a problem, because on-premise MySQL and PostgreSQL instances are still extremely common. Legacy applications that were never migrated. Databases that have to stay on-prem for data residency or latency reasons. Infrastructure acquired along with a company that hasn't been touched since. These instances don't disappear just because the rest of the stack moved to the cloud — they just become harder to see.
TL;DR
On-premise MySQL and PostgreSQL databases don't show up in any cloud API, so most "unified" multi-cloud tools quietly skip them. 1DataCloud connects to them directly via a read-only database connection and lists them in the same inventory as your AWS, GCP, and MongoDB Atlas instances — schema mapped, queryable in plain English. Try it free →
Why On-Premise Databases Become the Blind Spot
There's no cloud API to ask. AWS, GCP, and MongoDB Atlas all expose management APIs a tool can query to discover what exists. On-premise infrastructure has no equivalent — if a tool wants to know an on-prem database exists, someone has to tell it directly.
Discovery depends on tribal knowledge. On-prem inventory tends to live in people's heads, a wiki page, or a spreadsheet that was accurate as of some point last year. When the person who set up an instance leaves, the knowledge usually leaves with them.
It gets left out of "unified" views. Plenty of tools genuinely unify AWS and GCP, then stop there, because cloud APIs are the easy part. On-prem requires a direct connection to the database itself, so it's often skipped or bolted on as an afterthought.
Compliance treats it as a separate exercise. If cloud databases are inventoried automatically but on-prem ones are tracked manually, every audit ends up doing two different things — pull a report for the cloud side, then manually reconstruct the on-prem side.
A Better Workflow
Step 1 — Connect with a database connection string. Add the host, port, and a read-only database user for your on-prem MySQL or PostgreSQL instance. Since there's no cloud API, this is a direct database connection rather than an auto-discovery scan.
Step 2 — It shows up next to AWS, GCP, and Atlas. The on-prem instance appears in the same inventory list as your RDS, Cloud SQL, and Atlas clusters — engine, version, and status, side by side. One list, filterable by provider.
Step 3 — Schema loads automatically. The platform reads the database's tables, columns, data types, and relationships the same way it would for a cloud-hosted instance. No manual schema docs to maintain.
Step 4 — Query in plain English. With the schema mapped, Query1AI turns a plain-English question into read-only SQL for that on-prem instance, the same mechanism used for any connected database.
What's Actually Supported
On-premise MySQL, PostgreSQL, and MariaDB, via a direct read-only database connection — alongside the existing cloud coverage for AWS, GCP, and MongoDB Atlas, all in the same inventory. (Azure is still in active development and isn't part of this today.)
Where This Helps
One inventory instead of "cloud databases, plus that spreadsheet." Before: a dashboard for cloud databases and a separate, frequently stale document for whatever's still on-prem. After: everything's in the same list, filterable by provider.
Letting a non-technical teammate query a legacy on-prem database. Before: only the few engineers who remember the schema can get an answer out of it. After: anyone with dashboard access can ask a plain-English question through Query1AI once it's connected.
Onboarding an engineer to infrastructure that predates them. Before: they learn what's running on-prem by asking around. After: they see it in the inventory from day one, schema already mapped.
Security Model
- The database user should be read-only; queries Query1AI generates and executes are strictly read-only regardless — no INSERT, UPDATE, DELETE, or DDL
- The instance has to be network-reachable (firewall rule, VPN, or bastion) — nothing is installed on the database host itself
- Connection details are encrypted at rest, never shown in the UI after entry, never logged
- 1DataCloud stores schema metadata, not your data — query results return to your browser session and aren't retained server-side
1DataCloud for On-Premise Databases
- On-prem MySQL, PostgreSQL, and MariaDB listed alongside AWS, GCP, and MongoDB Atlas in one inventory
- Filter by provider to isolate what's running outside the cloud
- Schema metadata loaded automatically once connected
- Natural language querying via Query1AI, read-only
- One place to look, whether the answer lives on-prem or in the cloud
Free to try, setup in a few minutes.
👉 Connect your databases at 1datacloud.ai
Full write-up with FAQ and more detail on the security model: 1datacloud.ai/blog/on-premise-database-inventory-multi-cloud-view
Top comments (0)