[proxysql-cluster] Change suggestion to work more as normal charts

[tokheim]

Hi, I had a need for testing proxysql-cluster, but ended up doing quite a lot of changes to the chart. I'm open to splitting the changes up if you want to cherry-pick things etc. To highlight the features

• whole proxysql config can be configured through values.yaml
• • does not require modifying <chart>/files which I believe can't be done after chart packaging
• • with full flexibility, but still having convenient defaults
• • config set with yaml for easy overrides per environment
• set up github-pages as a chart repository, so you can run the chart without downloading source code
• config moved to secret since it contains passwords and such
• securitycontext set on the pod
• random admin password, password propagates to health checks and such without manually being kept in sync
• pod restart on config change through annotations
• poddisruptionbudget to avoid all pods being taken down during maintenances
• optional sidecar for tailing the eventslog data, so it can be picked up by more standard logging facilities
• optional sidecar "logrotating" the eventslog data, for pod not to run out of disk

The github action for packaging the chart requires all charts to be placed in a /helm folder, see helm/chart-releaser-action#78. To test the chart you can use

helm repo add testrepo https://tokheim.github.io/kubernetes
helm repo update
helm upgrade --install test-release testrepo/proxysql-cluster --debug --dry-run