A former Balena employee helped us get up and running OpenBalena 3.7.0 on Kubernets. However, we haven’t been impressed with the performance. It seems like things are just getting worse and worse with the K8 needing to be rebalanced on a daily basis at this point. We are not growing the fleet at this point.
Any ideas what could be causing this? Is there logs that need to be cleaned up? Any other maintenance that would help?
If not, is it possible to migrate the devices to a fresh setup?
I think this would depend a lot on how you packaged openBalena for kubernetes. I have been using these helm charts for openBalena on kubernetes, with my cluster running on GCP, and don’t seem to have any issues. It’s been stable with 3 control nodes and around 200 balena devices for the last year.
Do you have any CPU usage numbers for the different pods?
Our fleet is significantly larger. It was stable. Despite pausing fleet growth it’s getting worse and worse so looking for maintenance or migration ideas.