[00:00:00] So it appears it's not just me that's tracking what's going on in the world of sharding and Postgres, but apparently investors are looking at that too.
[00:00:11] So there were actually two investment announcements this week, both of them backing sharding projects. So we'll take a look at that. But I hope you, your friends, family and co workers continue to do well. Our first piece of content is our funding announcement. This is from pgdog.dev and if you've watched previous episodes, you know I've been watching what pgdog has been working on. They are developing essentially a Sharding pooler. So it's a product that exists outside of Postgres, it's not an extension, but it sits between your application and your databases and helps you shard them. And basically looking at this, they're talking about their funding announcement and he says, you know, Postgres is the only database you need and if you could just make it work with 100/t Terabyte table and a million queries per second, we don't think you'd need to use anything else. So they're just using the same old Postgres, just making it horizontally scalable and again, because it is independent of the database, you can deploy it anywhere on prem, in your cloud, etc. Now apparently from what they've deployed thus far, they're running about 2 million queries per second. I'm assuming that's across all their customers and they've sharded 20 terabytes they know about. The reason being is because PGDog is open source and anyone can just try it. They say they are coming out with a new version every week on Thursday and they have a discord community.
[00:01:36] So they are a three person startup still. But of course Lev ran Postgres at Instacart where they dealt with a lot of scaling issues in 2020. Now in terms of the investment, they raised 5.5 million from basis set, YC Pioneer Fund and other investors, basically giving them years of Runway.
[00:01:56] But if you want to learn more, definitely check out this blog post.
[00:02:00] Next one related to funding is Supabase Series F was
[email protected] and this round they say is for three things accelerate the development of open source and Postgres tools. And the number one thing they mention is their sharding solution Multigres, which is in 0.1 alpha now. So again this is equivalent to MySQL's the tests. So the person who initially designed it is now designing multigras for Postgres, the next thing they mentioned is to support their growth.
[00:02:30] So they do show an example of the growth here, which has been kind of crazy due to AI, apparently.
[00:02:36] So more than 60% of new databases are launched by some sort of AI tool, which is insane to me, but I guess those are the times we live in. And then the third they mentioned is liquidity for the employees.
[00:02:51] So definitely a lot of investors putting money behind this sharding Postgres area, which I find fascinating, but check it out if you want to learn more. Next piece of content how to test PostgreSQL 19 beta in your Kubernetes cluster this is from CloudNativePG IE Postgres 19 beta 1 was released last week and they're showing you how to get started with it on Kubernetes. So very brief but apparently easy to do, so you can check that out. Related to it beta testing PostgreSQL 19 with docker this is from andyatkinson.com and he's showing how you can get started with it on Docker. Now he did say that the official Postgres images for Docker are only for fully released versions, however JosephKit here this GitHub link has created a PR to add 19 beta 1 with instructions for how to use Docker build X and he followed these Instructions to get PG19 running on Docker and then he took it for a test drive. He checked out the new PGSTAT lock system view. He also checked out pgplannadvice and looked at a few different things with regard to PGSTAT statements as well and then of course tested out repack concurrently. So if you want to get started learning PG19 on Docker, definitely check out this piece of content.
[00:04:11] Next piece of content my three top PostgreSQL 19 features this is from Propensource at and her top three are the new option added to insert do select in addition to do nothing or do update. So she thought that was one of her top features. The second one repack and repack concurrently of course.
[00:04:33] And then the third is pgq, which is the Postgres graphed query language.
[00:04:39] And seeing this I thought what are my top three? So I actually took a look back at the release notes or the announcement and I'd say my top three are not having to manually adjust worker settings, have it dynamically choose the number of workers to use. Next is the two times better performance on inserts when foreign key checks are present. And I did look into this. There is a patch where there's a new fast path for doing foreign key checks and it essentially bypasses SPI or the server programming interface and uses low level index access functions as opposed to more like a query. But this enhancement sounds really exciting. And then for my third I think I have to choose that logical replication now replicate sequence values because every time I have to do logical replication upgrade you have to devise your own way to do it and having a built in way to do it is far better. But I will have to give an honorable mention to having repack and repack concurrently available. That's another great option for this release.
[00:05:48] Next Piece of Content Looking forward to PostgreSQL Query hints so this is from pgedge.com and this goes over pgplannadvice and has a lot of discussion about why have we not had pghints before which I've covered in previous episodes of scaling Postgres you can check out, but he takes pgplant advice through its paces actually try it out. So check this out if you want to learn more.
[00:06:13] Next piece of content A reviewer was born. This is from mydbanotebook.org and this is fascinating to me. So a friend of hers, Lucas, submitted his first patch to Postgres back in March that fixes a file descriptor leak when using IO method iouring and it has been sitting in the commit fest for two months with zero reviewers.
[00:06:36] And she says this is not unusual, this is the norm essentially because there are not that many committers and the patch queue is long. So her advice is if you submitted a patch and are waiting, the most useful thing you can do is not wait, but is to review someone else's patch.
[00:06:52] So in this blog post she walks through this particular patch step by step and sharing her findings and this is essentially her documenting how to do a real review.
[00:07:04] And the first thing that she mentions is that if you want your patch reviewed, review other patches. Because she says, quote, the community is not transactional about it, but the social dynamic is real. So basically if you make a name for yourself as a reviewer, people will notice when you then produce a patch. So she goes through finding a patch and assigning yourself to it, validating that the bug exists. So you replicate the bug first with the current build of Postgres, then you apply the patch, build it first, run the regression test to make sure nothing else is broken, and then you verify the bug is fixed so you're no longer seeing the behavior that existed, and then you review the patch from a code review perspective as well as a design perspective and then check any tests that were produced and write your review.
[00:07:54] So I think this is a great guide. Definitely encourage you to check it out Next Piece of content Acknowledged individuals in the PostgreSQL release notes 2026 edition. This is from markwkm.blogspot.com and apparently there are gifts that are sent to those who contribute directly to each new major release of Postgres. And he actually used where those gifts were going to see how different countries are contributing to Postgres and it's segmented by release and you can tell over the years. The individual years vary of course between countries, but over the years it looks like the US has contributed most, followed by Russia, followed by India, and then probably Germany, then Japan, China, France, etc. But he says you do need to take this with a little bit of a grain of salt because not everyone claims their gift, so it is a little underrepresented. So some of the totals may be off, but check this out if you're interested. And the last piece of content, how the other half counts. This is from thebuild.com and he's talking about analyze and how statistics are collected with Postgres and then he compares them to Oracle DB2, MySQL, SQLite, DuckDB, and Snowflake and rounding out with Postgres. So if you're interested in that, definitely check out this piece of content.
[00:09:22] I hope you enjoyed this episode. Be sure to check out scalingpostgres.com where you can find links to all the content discussed, as well as sign up to receive weekly notifications of each episode There. You can also find an audio version of the show as well as a full transcript. Thanks. I'll see you next week.
