In the last weeks I check multiple benchmarks and clusters in incidents. Often I had the feeling that the memory consumption from the gateway is not ideal or that there is a memory leak. I wanted to experiment regarding this memory consumptions. Since we saw in investigating https://github.com/zeebe-io/zeebe/issues/5641 a high memory spike when the gateway was not able to talk to other nodes I suspected that here might be some bugs hiding
Today I wanted to add new chaostoolkit experiment, which we can automate. We already have experiments like restarting followers and leaders for a partition, but in the past what also caused issues was multiple restarts/leader changes in a short period of time. This is the reason why I created #39.
First chaos day since my parental leave 馃帀.
Today I wanted to play a bit with ToxiProxy. Toxiproxy is a framework for simulating network conditions and ideal to do some chaos on the network.
In order to make our chaos experiments more realistic we have setup a new gke cluster, which is similar to the Camunda Cloud gke cluster. It allows us to test and experiment with Zeebe clusters which have the same configuration as Zeebe clusters in the Camunda cloud.
As part of the chaos day I run the same benchmark we normally run in our gke with our configuration against the Camunda Cloud Zeebe clusters.
Unfortunately I had no time today for new chaos experiment, but I spent with @pihme some time to investigate how we can run the cluster plans in our gke. We did a bit of progress. I'm finally able to create cluster plans in the ultratest and can extract all resource definitions via command line.
kubectl get pvc,configmap,service,deployment,statefulset,cronjob,storageclasses -o yaml --export | sed -e '/resourceVersion: "[0-9]\+"/d' -e '/uid: [a-z0-9-]\+/d' -e '/selfLink: [a-z0-9A-Z/]\+/d' -e '/status:/d' -e '/phase:/d' -e '/creationTimestamp:/d' > s-cluster.yaml
We now need to find a way to successfully deploy it in our cluster - it haven't been successful yet. We thought about using kustomize to adjust some values they use. Much easier would be to just deploy the operator they use in our gke cloud and use the CRD to deploy the cluster plans. I think we need to investigate a bit more here what is the best approach. In the end I would like to run our chaos experiments against clusters which correspond to the real deployed ones.