Use Redis Sentinel for Caching/Async Queries

Description

The PR https://github.com/stackabletech/superset-operator/pull/724 added celery worker support using Redis following the standard Superset example: https://superset.apache.org/admin-docs/configuration/async-queries-celery/

• This only works for RedisCache not RedisSentinelCache
• The schema used is always redis:// instead of sentinel://
• The master-name cannot be configured for sentinel nodes

So Redis in HA mode is not available without overrides (not tested).

Possible untested fix

configOverrides:
  superset_config.py:
    FILE_FOOTER: |
      from flask_caching.backends.rediscache import RedisSentinelCache
      _SENTINELS = [("redis-sentinel-0.redis-sentinel-headless", 26379),
                    ("redis-sentinel-1.redis-sentinel-headless", 26379),
                    ("redis-sentinel-2.redis-sentinel-headless", 26379)]
      _MASTER = "mymaster"
      _PW = os.environ["CELERY_RESULTS_BACKEND_DATABASE_PASSWORD"]

      RESULTS_BACKEND = RedisSentinelCache(
          sentinels=_SENTINELS, master=_MASTER, password=_PW, db=0,
          key_prefix="superset_results",
          # sentinel_password=..., only if the sentinels themselves require auth
      )

      _URL = ";".join(f"sentinel://:{_PW}@{h}:{p}" for h, p in _SENTINELS) + "/0"
      class CeleryConfig(object):
          broker_url = _URL
          result_backend = _URL
          broker_transport_options = {"master_name": _MASTER}
          result_backend_transport_options = {"master_name": _MASTER}
          imports = ("superset.sql_lab", "superset.tasks.scheduler")
          worker_prefetch_multiplier = 10
          task_acks_late = True
          task_annotations = {"sql_lab.get_sql_results": {"rate_limit": "100/s"}}
      CELERY_CONFIG = CeleryConfig

This is just a workaround and will need adaptation whenever the redis replicas change.

Issue checklist

This is a simple checklist of things to bear in mind when creating a new issue.

• Describe the use-case: As far as possible, use the pattern "As a [type of user], I would like [feature/functionality] to be able to do [specific action]." This helps identify the feature and the problem it addresses.
• Indicate importance and urgency: Use a scale (e.g., low, medium, high) to indicate the level of importance and urgency.
• Work-around: If there is a known work-around, describe it briefly.
• Environment: Describe the environment where the issue occurs (e.g., SDP version, K8S version, etc.).

Start by tracing the existing Redis and Celery configuration introduced by PR 724, including the superset_config.py override and CeleryConfig entry points. Compare it with RedisSentinelCache and the sentinel:// workaround shown here; done means Redis Sentinel can be configured for both caching and async-query broker/results without manual overrides, including the master name.