DBLab Engine v4.2.1 is out: fixes and security hardening, upgrade from v4.2.0 recommended.
- clone state saved atomically (a crash mid-write could destroy all clones)
- clone upgrade fixed for TLS-mounted setups
- safer upgrade image default (breaking)
gitlab.com/postgres-ai/databβ¦
Monitoring: We implemented a fix and are currently monitoring the result. postgresai.instatus.com/cmusβ¦
Investigating: Actualmente estamos investigando este incidente. postgresai.instatus.com/cmusβ¦
Investigating: Impact: All platform API requests are failing, so the console, CLI and
integrations are unusable. The website and marketing pages are unaffected. postgresai.instatus.com/cmurβ¦
DBLab by Postgres.AI ππ retweeted
DBLab Engine 4.2 is out!
Main thing: you can now pg_upgrade a clone with just one click or in CLI call:
==> dblab clone upgrade my-clone
=> automation stops the clone, "pg_upgrade --link" runs in place, then clone comes back on a new Postgres major version.
As usual, prod and every other clone untouched.
Didn't like the result?
==> dblab clone reset
No more excuses to avoid full-fledged testing before production Postgres gets upgraded.
Also in 4.2:
- paste a source URL, DBLab figures out provider, image, databases, settings
- branches and snapshots expire on policy (off by default)
- secrets from env vars (not plain-text in configs anymore)
- clone creation ~2x faster
DBLab by Postgres.AI ππ retweeted
traditional reminder that simple ALTER TABLE that felt instant on your laptop can put your production Postgres down
it needs to be cooked properly in production β use lock_timeout and retries postgres.ai/blog/20210923-zeβ¦
DBLab by Postgres.AI ππ retweeted
PgQue got a website with docs β pgque.dev
Identified: We are continuing to work on a fix for this incident. postgresai.instatus.com/cmpiβ¦
Investigating: We are currently investigating this incident. postgresai.instatus.com/cmpiβ¦
DBLab by Postgres.AI ππ retweeted
Working on this is a lot of fun. Here I shared how it all started, what original PgQ already has and what didn't, and what v0.2 brings (RC1 stamped yesterday)
New episode: "PgQue"
Nik and Michael discuss Nik's new project PgQue, a descendent of Skype's PgQ, for running queue-like workloads in Postgres.
π§ postgres.fm/episodes/pgque
πΊ youtu.be/bBNcR14Ts4U
DBLab by Postgres.AI ππ retweeted
What a nice blog post by Christophe Pettus
PgQue: Two Snapshots and a Diff thebuild.com/blog/2026/05/03β¦
DBLab by Postgres.AI ππ retweeted
Precise (not sampled) wait even analysis is something I didn't expect to have for Postgres - but here we go
Being able to zoom into concrete millisecond (!) and see the exact chain of wait events opens new possibilities for RCA, benchmarks, and performance studies
links π
New episode: "pg_wait_tracer"
Nik and Michael are joined by Dmitry Fomin to discuss his new tool pg_wait_tracer, as well as changes to core that would allow wait event tracing with lower overhead (and on managed services).
π§ postgres.fm/episodes/pg_waitβ¦
πΊ youtu.be/efY85jwfcQ8
DBLab by Postgres.AI ππ retweeted
New episode: "pg_wait_tracer"
Nik and Michael are joined by Dmitry Fomin to discuss his new tool pg_wait_tracer, as well as changes to core that would allow wait event tracing with lower overhead (and on managed services).
π§ postgres.fm/episodes/pg_waitβ¦
πΊ youtu.be/efY85jwfcQ8









