#has anyone done anything with profiling
1 messages · Page 1 of 1 (latest)
@fossil sail I did some manual benchmarking from compose + docker build to Dagger and in my case build time were a bit better in Dagger given that we were able to re-architect them more efficiently by using the SDKs. Happy to jump into a #911305510882513037 session tomorrow if you have time to help you understand what could be happening. I also have some experience with TestContainers as well
Could be an idea - I'm in +1200 (NZDT), not sure what timezone you're in, stuck in a bunch of meetings tomorrow.
One change here is going from an in-memory Java LDAP server to a container of OpenLDAP - I suspect setting up some TMPFS volumes for the data may be order.
Have you spotted what's particularly taking very long?
I'm currently in ART. starting my day here until 18/19hs
After switching my OpenLDAP service container to use tmpfs for its data my pipeline locally dropped back to 30 minutes, finally got it all running on AZDO and it seems to take about 10 minutes there - and I'm sure if/when I talk $boss into a Dagger Cloud account that'll drop.
And if I move that maven project to do less in a bigger reactor - could drop that further with cached steps.