[K8s / Devops] Add more configurables to Helm and publish the charts

Moved from GitHub dgraph/4386

Posted by hackintoshrao:

What version of Dgraph are you using?


Helm needs more options to configure the deployment, here is the list:

  • Configuration for anti-affinity between alphas and zero. This is required if the users want to schedule alphas and zeros on different nodes.

  • Configuration to set the cache size and remove the lru_mb option.

  • Publish the charts to the helm repo.

  • Configuration to enable Prometheus deployment, alerts, and scraping.

  • Configuration for node taints. This is required if the users don’t want to schedule any other pods on the nodes where the database servers are running.

  • Configuration to set both the initial resources to be allocated and resource limits for both alphas and zeros.

sleto-it commented :

@hackintoshrao @prashant-shahi - see also https://github.com/dgraph-io/dgraph/issues/4616

prashant-shahi commented :

@sleto-it I believe I can pick that task up. We can include this for the next sprint.

I found this in doing a search for charts. Assigning it to myself. A lot of these were already complete, so anything remaining, will run it past the finish line.