Database Performance
Key Performance Indicators - Response time, Errors, QPS, Threads running
Published: Tuesday, Aug 6, 2024 Last modified: Wednesday, Oct 2, 2024
AWS RDS
AWS focuses on underlying system resources which isn’t neccessarily the best way to monitor database performance.
Cloudwatch metrics:
- Query Response time = Query latency (Milliseconds) aka SelectLatency (note: Aurora only via AWS Performance Insights)
- Errors?
- QPS = Queries (Per second)
- Threads running = Connection utilization (Connections)
https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/cw-metrics-overview.html
- Response time: Corresponds to the “Query Latency” metric in CloudWatch, which measures the average time for read/write operations in milliseconds.
- Errors: Can be monitored through the MySQL error log, which is generated by default in RDS. The customer can enable exporting the logs to CloudWatch for easier monitoring and analysis.
- QPS (Queries Per Second): Corresponds to the “Queries” metric in CloudWatch.
- Threads running: Can be monitored through the “DatabaseConnections” metric in CloudWatch, which shows the number of client sessions connected to the DB instance. The customer can also use Enhanced Monitoring to see the list of OS processes and threads consuming CPU and memory resources.
Datadog Database Monitoring
- Response time = Select duration?
- Errors?
- QPS = Queries (Per second)
- Threads running = Active connections