Engineering blog · PG14 EOL positioning

Postgres 14 reaches end of life in 2026 — what now?

The Postgres community ends support for the 14.x series in September 2026. From that point on, self-hosted clusters running PG14 will not receive security patches, bug fixes, or compatibility updates from upstream — and the longer you stay on an unsupported major, the harder the eventual migration becomes. This explainer lays out the timeline, what “unsupported” actually means in practice, and how Veritas helps you cross the cliff.

Timeline
Key dates PG14 operators should plan around
  • September 2026 — Postgres 14 community EOL.Upstream postgresql.org stops publishing security patches and bug fixes for the 14.x series. Self-hosted clusters on PG14 become unsupported on that date.
  • Vendor-support cliffs begin.Managed-service SLAs and commercial vendor support tiers that track the community schedule start treating PG14 clusters as out-of-support. Some vendors extend paid backports; most don’t, and the ones that do charge a premium that grows with time.
  • Migration window opens — and narrows.The cleanest path is a planned major-version upgrade to PG16/PG17 over the next two release cycles. Every quarter past EOL compounds risk: unpatched CVEs, drift between client and server versions, and extensions that drop support first.
Where Veritas helps
Four engines, one recovery loop

Veritas is a multi-engine recovery layer that sits beside your existing operators and closes the loop on incident response — without replacing them and without paying a managed-service premium to keep PG14 alive.

  • Postgres
    launches with PG14 EOL recovery workflow
  • MySQL
    in preview
  • Cassandra
    in preview
  • Mongo
    in preview

The launch wedge is the PG14-EOL recovery workflow: detect that a self-hosted PG14 cluster is approaching the end of its support runway, generate a safe major-version upgrade plan, and execute it under a closed-loop recovery agent. MySQL, Cassandra, and Mongo ship preview builds so design-partner teams can validate the same agent pattern across their heterogeneous fleet.

Measured MTTH

What “recovery” actually means in our benchmark

Measured on a 3-node self-hosted Postgres cluster, single-node kill test, N=10 trials, median [TBD] seconds. Number lands with v0.2 alongside the first real-cluster rollout.

The benchmark methodology and full comparison matrix live on the home page under #mtth-benchmark; the same script is how we’ll measure Veritas on your cluster as a design partner.

Cross the PG14 cliff with us

We’re taking design partners for the PG14 recovery workflow ahead of the September 2026 EOL. Drop us a line and we’ll schedule a 30-minute working session on your cluster.

Become a PG14 EOL design partner

Or write directly to veritas-2-2@polsia.app.