PGConf.Russia 2021

In OZON, we work on PostgreSQL as a service. Currently, we have thousands of clusters. We know about the important questions software developers ask, and the most crucial ones are as follows: how to understand the database's workload and why this query is getting slow? In this presentation, I will explain how we help our developers answer these questions by simply having a look at our dashboard.

When working with PostgreSQL, we regularly encounter queries that take suspiciously long to be executed. What if it's too late to enable statistics or auto explain, and cancellation of the query doesn't look sensible, but you still need to find out what caused the issue or see at what stage execution the query is? pg_query_state is an open source PostgreSQL extension, which can give you access to the query execution plan in real time. It also makes available statistics for each node, including the number of processed rows.

We maintain large PostgreSQL databases with the total data volume exceeding 180 TB. Every instance in the cluster gets a workload of 15 thousand requests per second. These circumstances certainly impose limitations on the classical approach to updating database structures. The cost of a DBA error gets higher, as even a small error can lead to a sleepless night) In my talk, I will cover our existing limitations for changes in the production environment, non-classical DBMS behavior under heavy workloads, and will discuss PostgreSQL in general.

Crew Resource Management is a field which has become critical in many fields including aviation and nuclear power. However in IT, this field has not been well implemented. In this talk we will discuss what crew resource management is, what kinds of problems it solves, and why all database teams should implement it.

Several PostgreSQL-related case studies will be discussed and both technical and human factors will be part of the case studies.