I was unable to find something existing, so I'd love there to be a guide on how to effectively debug and analyze startup performance.
My use-case: We have a typical Quarkus application consisting of REST, hibernate-orm, some Kafka and some database. It has around 1k class files and 250 test class files in a single maven module.
We mostly use @QuarkusTests. Starting one takes over 30 seconds by now, and we'd like to improve that. Starting dev mode and hot-reloading has also become somewhat slow over time.
I have found out about quarkus.debug.print-startup-times in this issue, and I can try some regular profiling to try and figure out what takes how long.
But in case some people at Quarkus know how to really effectively and quickly determine what exactly takes how much time, and can name some Best Practices, it would be awesome to share that knowledge.
Description
I was unable to find something existing, so I'd love there to be a guide on how to effectively debug and analyze startup performance.
My use-case: We have a typical Quarkus application consisting of REST, hibernate-orm, some Kafka and some database. It has around 1k class files and 250 test class files in a single maven module.
We mostly use
@QuarkusTest
s. Starting one takes over 30 seconds by now, and we'd like to improve that. Starting dev mode and hot-reloading has also become somewhat slow over time.I have found out about
quarkus.debug.print-startup-times
in this issue, and I can try some regular profiling to try and figure out what takes how long.But in case some people at Quarkus know how to really effectively and quickly determine what exactly takes how much time, and can name some Best Practices, it would be awesome to share that knowledge.
Implementation ideas
No response