Personally I’m a big fan of email, just like blogging. To me a good email thread can be like a good novel where you’re following along always curious for what comes next. And no, I don’t mean the ones where there is an email to [email protected] and someone replies all, to only receive reply-all’s to not reply-all. I mean ones like started last week internally among the Azure Postgres team.

The first email was titled: Random Citus development and psql tips, and from there it piled on to be more and more tips and power user suggestions for Postgres. Some of these tips are relevant if you’re working directly on the Citus codebase, others relevant as anyone that works with Postgres, and some useful for debugging Postgres internals. While the thread is still ongoing here is just a few of the great tips:

Recently, I started working on the django-multitenant application. The main reason we created it was to to help django developers use citus in their app. While I was working on it, I wrote unit tests. And to be able to reproduce a customer’s production environment, I wanted the tests to use citus and not a single node postgres. If you are using citus as your production database, we encourage you to have it running in your development environment as well as your staging environments to be able to minimise the gap between dev and production. To understand better the importance of dev/prod parity, I recommend reading the Twelve-Factor app that will give you ideas to lower the chances of having last minute surprising when deploying on prod.

As part of the Citus team (Citus scales out Postgres horizontally, but that’s not all we work on), I’ve been working on pg_auto_failover for quite some time now and I’m excited that we have now introduced pgautofailover as Open Source, to give you automated failover and high availability!

When designing pg_auto_failover, our goal was this: to provide an easy to set up Business Continuity solution for Postgres that implements fault tolerance of any one node in the system. The documentation chapter about the pg_auto_failover architecture includes the following:

It is important to understand that pgautofailover is optimized for Business Continuity. In the event of losing a single node, then pgautofailover is capable of continuing the PostgreSQL service, and prevents any data loss when doing so, thanks to PostgreSQL Synchronous Replication.

Introduction to pg_auto_failover

The pg_auto_failover solution for Postgres is meant to provide an easy to setup and reliable automated failover solution. This solution includes software driven decision making for when to implement failover in production.

If you’ve worked with Ruby on Rails you likely have some understanding of how your database works with Rails, traditionally that has always meant specifying a single database per environment in your config/database.yml, possibly together with an environment setting like DATABASE_URL. Based on that configuration all reads and writes will access the database.

With Rails 6 this is about to change, thanks to the work of Eileen M. Uchitelle together with contributors from GitHub, Basecamp and Shopify. In the upcoming Rails 6 (currently in RC1), you will be able to easily change which database server you are connecting to, to support a variety of scenarios such as using read replicas and splitting your database into dedicated components.

The most interesting part, which we wanted to detail in this post, is related to configuring automatic queries against a read replicas, or follower database.

For roughly ten years now, I’ve had the pleasure of running and managing databases for people. In the early stages of building an application you move quickly, adding new tables and columns to your Postgres database to support new functionality. You move quickly, but you don’t worry too much because things are fast and responsive–largely because your data is small. Over time your application grows and matures. Your data model stabilizes, and you start to spend more time tuning and tweaking to ensure performance and stability stay where they need to. Eventually you get to the point where you miss the days of maintaining a small database, because life was easier then. Indexes were created quickly, joins were fast, count(*) didn’t bring your database to a screeching halt, and vacuum was not a regular part of your lunchtime conversation. As you continue to tweak and optimize the system, you know you need a plan for the future and know how you’re going to continue to scale.

Now in Preview: Introducing Hyperscale (Citus) on Azure Database for PostgreSQL

With Hyperscale (Citus) on Azure Database for PostgreSQL, we help many of those worries fade away. I am super excited to announce that Citus is now available on Microsoft Azure, as a new deployment option on the Azure Database for PostgreSQL called Hyperscale (Citus).

Hyperscale (Citus) scales out your data across multiple physical nodes, with the underlying data being sharded into much smaller bits. The same database sharding principles that work for Facebook and Google are baked right into the database. But, unlike traditional sharded systems, your application doesn’t have to learn how to shard the data. With Azure Database on PostgreSQL, Hyperscale (Citus) takes Postgres, the open source relational database, and extends it with low level internal hooks.

A few days ago a CVE was announced for Postgres. To say this CVE is a bit overblown is an understatement. The first thing to know is you’re likely completely safe. If you run on a managed service provider you are not going to be affected by this, and if you’re managing your own Postgres database all chances are you are equally as safe. This CVE received a note from Tom Lane on the pgsql-announce mailing list in response to it getting a broad amount of awareness and attention.

But, we thought this might be a good time to talk about a few principles and concepts that underly how Postgres works.

I talk with a lot of folks that set their database up, start working with it, and then are surprised by issues that suddenly crop up out of nowhere. The reality is, so many don’t want to have to be a DBA, instead you would rather build features and just have the database work. But your is that a database is a living breathing thing. As the data itself changes what is the right way to query and behave changes. Making sure your database is healthy and performing at it’s maximum level doesn’t require a giant overhaul constantly. In fact you can probably view it similar to how you approach personal health. Regular check-ups allow you to make small but important adjustments without having to make dramatic life altering changes to keep you on the right path.

After years of running and managing literally millions of Postgres databases, here’s my breakdown of what your regular Postgres health check should look like. Consider running this on a monthly basis to be able to make small tweaks and adjustments and avoid the drastic changes.

Choosing a database isn’t something you do every day. You generally choose it once for a project, then don’t look back. If you experience years of success with your application you one day have to migrate to a new database, but that occurs years down the line. In choosing a database there are a few key things to consider. Here is your checklist, and spoiler alert, Postgres checks out strongly in each of these categories.

SQL by itself is great and powerful, and Postgres supports a broad array of more modern SQL including things like window functions and common table expressions. But rarely do I write a query where I don’t want to tweak or format the data I’m getting back out of the database. Thankfully Postgres has a rich array of functions to help with converting or formatting data. These built-in functions save me from having to do the logic elsewhere or write my own functions, in other words I have to do less work because Postgres has already done it for me which I’m always happy about.

We’ve covered a set of functions earlier, today we’re going to look at some different categories of functions to dive deeper.

In an earlier blog post I wrote about how breaking problems down into a MapReduce style approach can give you much better performance. We’ve seen Citus is orders of magnitude faster than single node databases when we’re able to parallelize the workload across all the cores in a cluster. And while count (*) and avg is easy to break into smaller parts I immediately got the question what about count distinct, or the top from a list, or median?

Exact distinct count is admittedly harder to tackle, in a large distributed setup, because it requires a lot of data shuffling between nodes. Count distinct is indeed supported within Citus, but at times can be slow when dealing with especially larger datasets. Median across any moderate to large size dataset can become completely prohibitive for end users. Fortunately for nearly all of these there are approximation algorithms which provide close enough answers and do so with impressive performance characteristics.