Построение автоматического консультанта и инструментов настройки производительности в PostgreSQL

PostgreSQL is a mature and robust RDBMS since it has 30 years of history. Over the year, its query optimizer has been enhanced and usually produces good query plans.

However, can it always come up with good query plans? The optimization process has to use some assumptions to produce plans fast enough. Some of those assumptions are relatively easy to check (e.g. statistics are up-to-date), some harder (e.g. correct indexes are created), and some nearly impossible (e.g. making sure that the statistic samples are representative enough even for skewed data repartition). For now, given those various caveats, DBA sometimes can't always realize easily that they miss a chance to get a meaningful performance improvement.

To help DBA to get a truly good query plan, we'll present below some tools that can help to fix some of those problems by providing a missing index adviser, looking for extended statistics to create, and row estimation error correction information to get appropriate join orders with join methods automatically.

• pg_qualstats: provides a new index and extended statistics suggestions to gather many predicate statistics on the production workload.
• pg_plan_advsr: provides alternative good query plans automatically to analyze iterative query executions information to fix estimation rows error.

In this talk, we will explain how those tools work under the hood and see what can be done, how they can work together. Also, we will mention what other tools also exist for related problems. Therefore, it will be useful for DBA who are interested in improving query performance or want to check whether current settings of indexes and statistics are adequate.

Продвинутые техники pg_upgrade

When it comes to performing postgres upgrades between major versions the command line utility pg_upgrade is the most popular method used today. But as with all things wonderful, there are caveats. One of the more critical issues is what to do when there's a failure. The purpose of this talk is to present those little covered "tricks" of the trade that allows one to improve and enhance the upgrade experience.

We'll start out by reviewing its basic mode operation. We'll then learn what makes it so darn fast when upgrading multi-terabyte datacluster in a matter of minutes. Finally, you will be introduced to those feared failures and how to deal with them with confidence and certainty.

Herein is a summary of the topics in this talk:

• How pg_upgrade works: the big picture
• About pgupgrade (comand line invocation)
  • arguments and options

• The steps performing an upgrade
• About the replication ROLE
  • with replication
  • with login

• Upgrade options: COPY vs Hard Link
• Post-upgrade
  • about performance
  • aboutanalyze
  • repack
  • reindexing

• When something goes wrong at the point of no return (or is it?)
• Upgrading the REPLICA
  • default method: pg_basebackup
  • the cool method:
    • - leveraging rsync
    caveat; about vacuum wraparound

Архитектуры с Postgres в продакшне

When using PostgreSQL in production it is important to implement a strategy for High Availability. With a database service, the HA properties apply to both the service itself and of course to the data set.

In this talk we learn how to think about specific HA needs of your production environment and how to achieve your requirements with Open Source tooling when it comes to your database of choice, PostgreSQL. In particular, we dive in many options that could be implemented for Postgres to evolve its offering from being a toolset to being “batteries included”. What does it mean in the context of HA? How to achieve it?

Как преобразовать Postgres в облачную платформу

Is deploying Postgres in Kubernetes just repackaging it into a container? Can’t Postgres leverage the wide range of Cloud-Native software and integrate well with K8s? Join this journey that will cover and demonstrate, with demos running on StackGres:

• How to structure Postgres into an init-less container, plus several sidecar containers for connection pooling, backups, agents, etc.
• Defining high level CRDs as the single API to interact with the Postgres operator.
• Using K8s RBAC for user authentication of a web UI management interface.
• Using Prometheus for monitoring; bundling a node, Postgres and PgBouncer exporters together.
• Proxying Postgres traffic through Envoy. Terminate Postgres SSL with an Envoy plugin, that also exports wire protocol metrics to Prometheus.
• Using Fluentbit to capture Postgres logs and forward them to Fluentd, which stores them on a centralized Postgres database.

You will be able to follow the session on your own Kubernetes cluster, and go from zero to a Postgres hero on Kubernetes with little effort! Create in minutes your own Postgres-as-a-Service on your Kubernetes.