After `MESSAGE_BUS_REDIS_URL` was introduced, the HA template fails to deploy due to missing corresponding parameter.

In the past, system component used to inherit the connection string set for `REDIS_URL` and then changed the logical db to `/8`. This second connection was used for message bus. Now, if not specified, system assumes `redis://system-redis/8` as default[1], which is not a valid option in HA deployments. As a consequence, system-(app|sidekiq|sphinx) keeps crashing until `MESSAGE_BUS_REDIS_URL` env var is set.

[1] https://github.com/3scale/porta/blob/734d23e1bf53139fe0a071e3f6057420b9753a99/openshift/system/config/message_bus.yml#L4

We basically have 3 different options to address this issue:

1. Document the workaround (i.e. how to manually set the missing env vars) and users will have to manually define a new env var to get the HA deploy profile working
   This is pretty hacky and not nice, as people won't read the docs until it fails.
2. Change to HA template (and Operator HA deploy) requiring a new release of both, and some new testing.
   We see long-term maintainability issues with this, and upgrade complexity, in Operator version in particular.
3. Code change to system to avoid the issue and a re-release and some re-testing of System by QE
   Probably the best fix in terms of customer impact and longer-term maintainability.....but needs a new System release.