Description of problem:

Customer has over 2400 Devspace installed on a single cluster.

The cluster is slow and irresponsive periodically. 

We found that etcd is nearly filing up (7.5 GB) and configmap take around 5.5 GB.

The largest configmaps are ca-cert-merged and che-trusted-ca-cert which are 1 MB each and they can be found in every devspace namespace.

The content of ca-cert-merged configmap in different namespaces are actually same.

This make kube-apiserver request are slow and take a lot of memory.

The cluster is running on 4.16.35.

Actual results:

Cluster is slow and irresponsive periodically

Expected results:

Cluster should be functional and stable