We use the provider to create Dashboards, Folders, etc.
What we observe is an incredible high amount of CPU usage.
We have around 100 - 400 different resources per cluster and the uses the cpu of a complete 8vCPU core node in our cluster the whole time. Changes to a single dashboard can take a long time until reconciled.
We use the provider to create Dashboards, Folders, etc.
What we observe is an incredible high amount of CPU usage. We have around 100 - 400 different resources per cluster and the uses the cpu of a complete 8vCPU core node in our cluster the whole time. Changes to a single dashboard can take a long time until reconciled.
It feels connected to this: https://github.com/upbound/upjet/issues/116
We tried to configure the provider with
--sync=24h
but that did not change anything.I would think it is related to the
PollIntervall
setting ( https://github.com/grafana/crossplane-provider-grafana/blob/4644dab3fdf3554ac31cd9f212c7b8c9d524b60b/cmd/provider/main.go#L80 ).The AWS provider allows a configuration of the interval (see https://github.com/upbound/provider-aws/pull/140/files ).
I would propose we make the interval configurable for this provider as well and try to provide a patch for that.