Bitbucket runner with autoscaler incredible slow

[Fabian Börner]

Hi,

our bitbucket runner in kubernetes runs incredible slow, for builds that take locally less then 5 minutes it takes up to 90 minutes to complete the build.

I have already set high memory and cpu limits but this is the consumption of the runner

kubectl -n bitbucket-runner-control-plane top pod runner-controller-556f77fdb5-9hj8q
NAME CPU(cores) MEMORY(bytes)
runner-controller-556f77fdb5-9hj8q 0m 53Mi

it barely uses the cpu and only consumes 54Mi sometimes 500 but not more even if the request assigned is 8 GB

Someone has an idea how to resolve this?

[Gajesh Bhat]

Have you tried deleting the existing pods so that the new changes take effect? I had a similar problem, and increasing the disk space of the VMs helped fix the issue. It can be a bit slow when new Runners are being provisioned, but it shouldn't be as slow as you’re experiencing.

Make sure to:

1. Delete existing pods to apply new resource configurations:

sudo kubectl delete pod -l app=runner-controller -n bitbucket-runner-control-plane

2. Check and increase disk space if necessary:

df -h

Other Possible solutions you can explore are:

1. Pod Affinity and Node Selection (Multi-region Cloud ) : Ensure pods are scheduled on nodes with adequate resources and minimal contention.

2. Network Latency and Bandwidth : Ensure optimal network performance, as slow network connections can impact build times.