Query Throughput
Query throughput is the number of operations (select or commit) your database is handling each second. If you are experiencing anomalous query throughput, check the code that accesses the database. This can be done by:
- Correlating the anomaly with a release of new code
- Checking the database logs in the RDS console for the offending query
Query Latency
Database query latency can cause big problems for your application, and indicates that your queries are slower. It can occur for one of two reasons:
- Your application code is sending more queries to your database than it can handle
- Underlying hardware issues
It is more likely that the latency is occurring due to your application code. Following the advice above for Throughput anomalies should help you find the offending code.
Once you’ve determined that your application code is not to blame for the increased latency, you can try to solve the problem of underlying hardware issues. You can either:
- Add read replicas to an Aurora cluster to help with select latency
- Submit a support request to AWS
- Move your database to a different physical host by either initiating a manual failover or by scaling up your instance
Resources
