Helm: Remove graphite-web from graphite proxy configuration

[fionaliao]

What this PR does

This PR removes the graphite-web component from the graphite proxy.

graphite-web is used to render queries that the native query engine cannot. However, the Helm chart configuration for graphite-web is broken so it's not been running and handling queries.

Issues:

• The graphite-web container has to writes to files in the /opt/ and /var/ directories and also changes the ownership of some /opt/ files. It was unable to start up due to the restricted policy that was applied to the pod/container.
• graphite-web is hardcoded to listen on port 80 (code). In the Helm chart it's configured to use on port 8080, so even without the restricted policy, we were failing to query it.
• graphite-web expects the host header for the request to be graphite (code). This requires the graphite-web service name to be just graphite instead of {{ include "mimir.resourceName" (dict "ctx" . "component" "graphite-web") }}.

We could either fix these issues or remove graphite-web completely. I decided to remove graphite-web, as the majority of graphite functions and features are supported by the native GEM graphite query engine, and we're currently working on supporting the remaining few issues. Also the configuration for graphite-web has never worked since it was added to the Helm chart, so we're not removing any features. The configuration options for graphite-web are still available in GEM if any users want to use it.

Testing

I tested this manually by deploying the Helm chart locally in a Kind cluster and verifying a) the graphite-web resources had been removed, and b) all the other pods were successfully running. Also ran the test_graphite_write_path.sh and test_graphite_read_path.sh scripts, checked they worked, and also that queries that used to fallback to graphite-web were returning with the expected errors.

Checklist

• Tests updated
• Documentation added
• CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX]

[dimitarvdimitrov]

LGTM, I don't have a concern about introducing a breaking change without deprecation since this hasn't been working for a long time now.