I've recently faced frequent timeouts for Sequelize queries running in production against a Postgres DB. Such timeouts were caused by a moderate requests load. I aim to understand
This little project is sort of a test-bench that I'll use to play with load-testing Sequelize using k6.
Useful references:
- https://www.theserverside.com/blog/Coffee-Talk-Java-News-Stories-and-Opinions/How-to-setup-an-Nginx-load-balancer-example
- https://richyen.com/postgres/2021/09/03/less-is-more-max-connections.html
- https://github.com/brettwooldridge/HikariCP/wiki/About-Pool-Sizing
- https://stackoverflow.com/questions/2757549/org-postgresql-util-psqlexception-fatal-sorry-too-many-clients-already
- Postgres has the configuration option
max_connections. Refer to this blog post for a nuanced explanation of how to reason about this variable. In brief, beyond the value of a couple hundred, you should consider using a connection pooler like PgBouncer.
docker run --rm -i --network="host" grafana/k6 run - <./k6/script.js
docker run -i --rm --network="host" ghcr.io/szkiba/xk6-dashboard:latest run --out=dashboard - <./k6/script.js