Abstract:
In database monitoring, more metrics don’t always mean better insights. This talk cuts through the complexity to reveal the essential metrics that truly matter when monitoring MySQL and PostgreSQL. Through real-world examples, we’ll demonstrate how focusing on fewer, but more meaningful metrics leads to better database reliability.
We’ll explore key monitoring principles through practical examples from both databases:
MySQL Focus:
– How to effectively monitor thread states to identify blocked queries and resource contention
– Why tracking InnoDB row lock wait times and lock wait operations provides better insights than watching generic lock metrics
– The power of combining threads_running and slow query analysis instead of drowning in status variables
PostgreSQL Focus:
– Understanding pg_stat_activity wait_events to spot bottlenecks and blocking queries
– The three critical autovacuum metrics that matter more than tracking every vacuum parameter
– Using state transitions in pg_stat_activity (active, idle, idle in transaction) to identify performance patterns
Shared Principles:
– A simple approach to query latency monitoring that works for both databases
– Universal techniques for identifying troubled queries using execution time and frequency
– How to consolidate monitoring between MySQL and PostgreSQL instances for teams running both
– Smart alerting strategies that prevent alert fatigue while catching real problems
Perfect for developers and database administrators who want to build effective monitoring without the complexity. You’ll leave understanding how to implement a minimalist but powerful monitoring strategy that catches real problems while reducing alert fatigue, with clear guidance on which approaches to use for each database system.