Nik Samokhvalov
@samokhvalov.bsky.social
Talk to me if you have issues with your Postgres databases: nik@postgres.ai // I stand with Ukraine // Postgres.AI & DBLab & Postgres.FM
postges_ai monitoring 0.12: SLRU, MultiXact member space exhaustion, many improvements and fixes
GitHub: github.com/postgres-ai/...
GitLab: gitlab.com/postgres-ai/...
GitHub: github.com/postgres-ai/...
GitLab: gitlab.com/postgres-ai/...
Releases · postgres-ai/postgres_ai
Expert-level Postgres monitoring tool designed for humans and AI systems - postgres-ai/postgres_ai
github.com
November 11, 2025 at 7:03 PM
postges_ai monitoring 0.12: SLRU, MultiXact member space exhaustion, many improvements and fixes
GitHub: github.com/postgres-ai/...
GitLab: gitlab.com/postgres-ai/...
GitHub: github.com/postgres-ai/...
GitLab: gitlab.com/postgres-ai/...
November 11, 2025 at 6:58 PM
See it all the time: developers want faster DBs but only add indexes – never drop unused or rebuild degraded ones.
This is why postgres.ai is powerful: it detects indexes that need to be dropped, sets up automated reindex, and observes the results.
So you can work on features, not DB maintenance.
This is why postgres.ai is powerful: it detects indexes that need to be dropped, sets up automated reindex, and observes the results.
So you can work on features, not DB maintenance.
PostgresAI copilot — Expert-level Postgres observability | PostgresAI
postgres.ai
November 11, 2025 at 12:44 AM
See it all the time: developers want faster DBs but only add indexes – never drop unused or rebuild degraded ones.
This is why postgres.ai is powerful: it detects indexes that need to be dropped, sets up automated reindex, and observes the results.
So you can work on features, not DB maintenance.
This is why postgres.ai is powerful: it detects indexes that need to be dropped, sets up automated reindex, and observes the results.
So you can work on features, not DB maintenance.
#PostgresMarathon 2-012: Ultra-fast replica creation with pgBackRest postgres.ai/blog/2025110...
Couldn't believe the numbers: 35 TiB/h copying PGDATA from one EC2 instance to another.
That's fast enough to clone a 10 TiB production database in ~17 minutes.
1/
Couldn't believe the numbers: 35 TiB/h copying PGDATA from one EC2 instance to another.
That's fast enough to clone a 10 TiB production database in ~17 minutes.
1/
November 6, 2025 at 6:09 AM
#PostgresMarathon 2-012: Ultra-fast replica creation with pgBackRest postgres.ai/blog/2025110...
Couldn't believe the numbers: 35 TiB/h copying PGDATA from one EC2 instance to another.
That's fast enough to clone a 10 TiB production database in ~17 minutes.
1/
Couldn't believe the numbers: 35 TiB/h copying PGDATA from one EC2 instance to another.
That's fast enough to clone a 10 TiB production database in ~17 minutes.
1/
Redesigning @postgres_ai on the path to Self-Driving Postgres.
First: tackling the problem most teams actually face: index maintenance, not index creation.
Sad truth: everyone agrees when we show what needs to be done. But it's too much effort, so it sits in TODO forever. Until DB hits a wall
1/
First: tackling the problem most teams actually face: index maintenance, not index creation.
Sad truth: everyone agrees when we show what needs to be done. But it's too much effort, so it sits in TODO forever. Until DB hits a wall
1/
November 6, 2025 at 4:29 AM
Redesigning @postgres_ai on the path to Self-Driving Postgres.
First: tackling the problem most teams actually face: index maintenance, not index creation.
Sad truth: everyone agrees when we show what needs to be done. But it's too much effort, so it sits in TODO forever. Until DB hits a wall
1/
First: tackling the problem most teams actually face: index maintenance, not index creation.
Sad truth: everyone agrees when we show what needs to be done. But it's too much effort, so it sits in TODO forever. Until DB hits a wall
1/
there are performance cliffs in Postgres, there are just cliffs -- they are not less painful to fall from
excited about new postgres.ai product we're building now – it will help predict / detect / mitigate all possible cliffs
excited about new postgres.ai product we're building now – it will help predict / detect / mitigate all possible cliffs
November 3, 2025 at 11:06 PM
there are performance cliffs in Postgres, there are just cliffs -- they are not less painful to fall from
excited about new postgres.ai product we're building now – it will help predict / detect / mitigate all possible cliffs
excited about new postgres.ai product we're building now – it will help predict / detect / mitigate all possible cliffs
Deep dive in the situation with prepared statements and partitioned tables in Postgres:
part 1 postgres.ai/blog/2025102...
part 2 postgres.ai/blog/2025102...
part 3 postgres.ai/blog/2025103...
part 1 postgres.ai/blog/2025102...
part 2 postgres.ai/blog/2025102...
part 3 postgres.ai/blog/2025103...
#PostgresMarathon 2-011: Prepared statements and partitioned tables — the paradox, part 3 | PostgresAI
In #PostgresMarathon 2-009 and #PostgresMarathon 2-010, we explored why execution 6 causes a lock explosion when building a generic plan for partitioned tables — the planner must lock all 52 relations...
postgres.ai
October 31, 2025 at 8:14 PM
Deep dive in the situation with prepared statements and partitioned tables in Postgres:
part 1 postgres.ai/blog/2025102...
part 2 postgres.ai/blog/2025102...
part 3 postgres.ai/blog/2025103...
part 1 postgres.ai/blog/2025102...
part 2 postgres.ai/blog/2025102...
part 3 postgres.ai/blog/2025103...
this repo deserves more stars gitlab.com/postgres-ai/...
we aim to have most detailed and advanced monitoring for Postgres -- and it's FOSS because it's just a foundation of something bigger
(after 20 years with Postgres, next month will be the most exciting announcement of my career)
we aim to have most detailed and advanced monitoring for Postgres -- and it's FOSS because it's just a foundation of something bigger
(after 20 years with Postgres, next month will be the most exciting announcement of my career)
October 31, 2025 at 8:13 PM
this repo deserves more stars gitlab.com/postgres-ai/...
we aim to have most detailed and advanced monitoring for Postgres -- and it's FOSS because it's just a foundation of something bigger
(after 20 years with Postgres, next month will be the most exciting announcement of my career)
we aim to have most detailed and advanced monitoring for Postgres -- and it's FOSS because it's just a foundation of something bigger
(after 20 years with Postgres, next month will be the most exciting announcement of my career)
postgres_ai monitoring 0.11: VictoriaMetrics, lots of optimizations, new dashboards for table and index analysis, and more github.com/postgres-ai/...
Releases · postgres-ai/postgres_ai
Expert-level Postgres monitoring tool designed for humans and AI systems - postgres-ai/postgres_ai
github.com
October 31, 2025 at 8:12 PM
postgres_ai monitoring 0.11: VictoriaMetrics, lots of optimizations, new dashboards for table and index analysis, and more github.com/postgres-ai/...
postgres_ai monitoring 0.11: VictoriaMetrics, lots of optimizations, new dashboards for table and index analysis, and more github.com/postgres-ai/...
Releases · postgres-ai/postgres_ai
Expert-level Postgres monitoring tool designed for humans and AI systems - postgres-ai/postgres_ai
github.com
October 30, 2025 at 9:26 PM
postgres_ai monitoring 0.11: VictoriaMetrics, lots of optimizations, new dashboards for table and index analysis, and more github.com/postgres-ai/...
TIL during this hacking session: REFRESH MATERIALIZED VIEW CONCURRENTLY creates temp tables github.com/postgres/pos... – so, if you do it super frequently, it might contribute to system catalog bloat
October 29, 2025 at 6:01 PM
TIL during this hacking session: REFRESH MATERIALIZED VIEW CONCURRENTLY creates temp tables github.com/postgres/pos... – so, if you do it super frequently, it might contribute to system catalog bloat
live session of Postgres [vibe] hacking happening now -- join! www.youtube.com/watch?v=6cZv...
[Vibe] Hacking Postgres with Andrey, Kirk, and Nik
YouTube video by PostgresTV
www.youtube.com
October 29, 2025 at 4:46 PM
live session of Postgres [vibe] hacking happening now -- join! www.youtube.com/watch?v=6cZv...
I'm on a trip until 2025-10-22, and have only limited time with laptop – #PostgresMarathon will be continued when I'm back
October 16, 2025 at 4:33 AM
I'm on a trip until 2025-10-22, and have only limited time with laptop – #PostgresMarathon will be continued when I'm back
#PostgresMarathon 2-008: LWLock:LockManager and prepared statements postgres.ai/blog/2025101...
#PostgresMarathon 2-008: LWLock:LockManager and prepared statements | PostgresAI
As was discussed in #PostgresMarathon 2-002, for a simple SELECT from a table, at planning time, Postgres locks the table and all of its indexes with AccessShareLock. A simple demo to remind it (let m...
postgres.ai
October 15, 2025 at 1:16 PM
#PostgresMarathon 2-008: LWLock:LockManager and prepared statements postgres.ai/blog/2025101...
#PostgresMarathon 2-007: Should we worry about pg_blocking_pids()'s observer effect? postgres.ai/blog/2025101...
#PostgresMarathon 2-007: Should we worry about pg_blocking_pids()'s observer effect? | PostgresAI
Many years ago, when developing complex automated procedures for a large company, I realized that my automation needs monitoring components. Including understanding heavyweight lock contention – for e...
postgres.ai
October 14, 2025 at 12:31 AM
#PostgresMarathon 2-007: Should we worry about pg_blocking_pids()'s observer effect? postgres.ai/blog/2025101...
Got very much attracted by title "Boosting Planning Performance | Scaling Postgres 387",
was very curious (because I do want to learn more ways to boost planning performance), checked it out: www.youtube.com/watch?v=d7__...
...and ERROR: infinite recursion detected
was very curious (because I do want to learn more ways to boost planning performance), checked it out: www.youtube.com/watch?v=d7__...
...and ERROR: infinite recursion detected
Boosting Planning Performance | Scaling Postgres 387
YouTube video by Scaling Postgres
www.youtube.com
October 12, 2025 at 11:40 PM
Got very much attracted by title "Boosting Planning Performance | Scaling Postgres 387",
was very curious (because I do want to learn more ways to boost planning performance), checked it out: www.youtube.com/watch?v=d7__...
...and ERROR: infinite recursion detected
was very curious (because I do want to learn more ways to boost planning performance), checked it out: www.youtube.com/watch?v=d7__...
...and ERROR: infinite recursion detected
Postgres hacking session LIVE www.youtube.com/watch?v=5lN7...
[Vibe] Hacking Postgres with Andrey, Kirk, Nik
YouTube video by PostgresTV
www.youtube.com
October 8, 2025 at 5:08 PM
Postgres hacking session LIVE www.youtube.com/watch?v=5lN7...
#PostgresMarathon 2-001
To warm up, let's talk about lightweight and heavyweight locks (or "regular locks" or just "locks").
I'm using these materials:
- PG docs (first of all, www.postgresql.org/docs/current...)
1/
To warm up, let's talk about lightweight and heavyweight locks (or "regular locks" or just "locks").
I'm using these materials:
- PG docs (first of all, www.postgresql.org/docs/current...)
1/
13.3. Explicit Locking
13.3. Explicit Locking # 13.3.1. Table-Level Locks 13.3.2. Row-Level Locks 13.3.3. Page-Level Locks 13.3.4. Deadlocks 13.3.5. Advisory Locks PostgreSQL provides various …
www.postgresql.org
October 5, 2025 at 6:55 AM
#PostgresMarathon 2-001
To warm up, let's talk about lightweight and heavyweight locks (or "regular locks" or just "locks").
I'm using these materials:
- PG docs (first of all, www.postgresql.org/docs/current...)
1/
To warm up, let's talk about lightweight and heavyweight locks (or "regular locks" or just "locks").
I'm using these materials:
- PG docs (first of all, www.postgresql.org/docs/current...)
1/
#PostgresMarathon is back!
Make sure you follow me so we could learn some good stuff about Postgres together.
This time, I'll be doing it a bit differently:
- I'll choose one big Postgres topic that I'll be covering over a week or two
1/
Make sure you follow me so we could learn some good stuff about Postgres together.
This time, I'll be doing it a bit differently:
- I'll choose one big Postgres topic that I'll be covering over a week or two
1/
October 4, 2025 at 1:45 AM
#PostgresMarathon is back!
Make sure you follow me so we could learn some good stuff about Postgres together.
This time, I'll be doing it a bit differently:
- I'll choose one big Postgres topic that I'll be covering over a week or two
1/
Make sure you follow me so we could learn some good stuff about Postgres together.
This time, I'll be doing it a bit differently:
- I'll choose one big Postgres topic that I'll be covering over a week or two
1/
anti-tip:
SELECT INTO is ...DDL. It creates a table.
Not what you expect after spending hours inside pl/pgsql
www.postgresql.org/docs/current...
I wish Postgres didn't have this feature 😂
SELECT INTO is ...DDL. It creates a table.
Not what you expect after spending hours inside pl/pgsql
www.postgresql.org/docs/current...
I wish Postgres didn't have this feature 😂
SELECT INTO
SELECT INTO SELECT INTO — define a new table from the results of a query Synopsis [ WITH [ RECURSIVE …
www.postgresql.org
October 4, 2025 at 12:29 AM
anti-tip:
SELECT INTO is ...DDL. It creates a table.
Not what you expect after spending hours inside pl/pgsql
www.postgresql.org/docs/current...
I wish Postgres didn't have this feature 😂
SELECT INTO is ...DDL. It creates a table.
Not what you expect after spending hours inside pl/pgsql
www.postgresql.org/docs/current...
I wish Postgres didn't have this feature 😂
If you think that you understand the whole meaning of max_locks_per_transaction, think again
October 3, 2025 at 8:31 PM
If you think that you understand the whole meaning of max_locks_per_transaction, think again
What do you like most in Postgres 18? Look at www.postgresql.org/docs/release... and tell me.
I'll go first, my top-3 list:
I'll go first, my top-3 list:
PostgreSQL: Release Notes
www.postgresql.org
September 30, 2025 at 7:27 PM
What do you like most in Postgres 18? Look at www.postgresql.org/docs/release... and tell me.
I'll go first, my top-3 list:
I'll go first, my top-3 list:
Some PG18 posts spread misinformation.. Be careful, we see a lot of LLM-generated BS lately.
September 25, 2025 at 3:55 PM
Some PG18 posts spread misinformation.. Be careful, we see a lot of LLM-generated BS lately.
If you deal with index bloat, and ready to try out a new thing – reach out to me, we're have something really cool
September 24, 2025 at 6:21 PM
If you deal with index bloat, and ready to try out a new thing – reach out to me, we're have something really cool