Schema auto-management support for PostgreSQL 13+ in self-hosted Temporal

Hi, we are experimenting adding temporal to our stack…

Environment:

Summary:
We’re running self-hosted Temporal against PostgreSQL 18 (Cloud SQL). Based on the current docs and our testing, it appears the only officially supported plugin for managed schema (manageSchema: true) is postgres12, which is deprecated. For any PostgreSQL version newer than 12, schema setup and upgrades have to be performed manually using temporal-sql-tool. We’d like to confirm this is correct and understand the roadmap, because the manual upgrade path adds significant operational overhead for us.

What we tried:

  1. pluginName: postgres12 with manageSchema: true against PostgreSQL 18 — failed. The database initialization errored out and schema could not be created properly.

  2. Working configuration we ended up with:

    manageSchema: false
    pluginName: postgres12_pgx
    driverName: postgres12_pgx
    

    With this config we had to:

    • Manually create the users and databases in PG18
    • Run temporal-sql-tool against each database to set up the schema
    • Only then did Temporal start successfully

Reference docs we used:
Upgrade the Temporal Server | Temporal Platform Documentation — confirms that for self-hosted PostgreSQL the schema upgrade path is via the CLI tool, with no auto-upgrade.

Operational concern:
Because schema is not auto-managed on PG13+, every Temporal server upgrade that ships a schema change requires us to:

  1. Scale Temporal down
  2. Run the schema upgrade manually via temporal-sql-tool
  3. Deploy the new Temporal version

This is workable but manual and error-prone, and the Helm chart doesn’t help here — it only assists with the initial install, not ongoing schema upgrades.

  1. Is there a supported plugin for managed schema on PostgreSQL 13+ (and specifically PG18) that we’re missing? If postgres12 is expected to work against newer PG versions, what’s the supported configuration?
  2. Is automatic schema management for newer PostgreSQL versions on the roadmap? If so, is there a timeline?
  3. Is there a recommended pattern for automating the scale-down → schema upgrade → redeploy flow in a Helm/Kubernetes environment, or any tooling Temporal recommends to reduce the manual steps?
  4. Are there any known issues or gotchas with running Temporal against Cloud SQL PostgreSQL 18 specifically that we should be aware of?

Happy to share configs, logs from the failed manageSchema: true attempt, or anything else that would help.

Thanks in advance for your help,