Quick Wins for Performance Troubleshooting With pg stat monitor - Pablo Svampa, Percona

Think about underperforming queries, unstable execution plans, IO spikes, and fluctuating workloads; some of the many common issues on the database's ground.
Now, think about those many times they took place in the middle of the night or when they came up in random and short bursts, and you were not in front of your computer.
Troubleshooting performance issues that occurred in the past and writing RCA reports requested by upper management could be challenging. It becomes a headache if we lack a powerful monitoring tool or database-session activity data.
What if you could quickly review DDL and DML activity in the past? What if you could quickly identify not only the most expensive queries but their execution plans, IO requests, CPU, and more?
This session will cover how to quickly set up, collect, analyze data, and identify issues that took place in the past without adding the need for a monitoring tool and additional HW resources.
Spot out IO spikes, CPU spikes, unstable execution plans, load fluctuations, and more with pg_stat_monitor and some custom SQL.
The approach will help you collect metrics and troubleshoot issues while deciding on your following monitoring tool and stepping forward.
#perconalive #postgresql