Introduce background mechanism to retry enablement of Redis related features

Context

Outcome from this thread: !2131 (comment 2380954849).

Problem

Right now we lack the ability to self-heal and enable features that depend on Redis and have been disabled/skipped due to a failure to connect to Redis at startup.

Goal

• Enable features to self-heal/enable themselves in the background in case Redis becomes available post startup;
• Decrease Redis connection timeout during startup as it can delay the service start for too long and will be unnecessary given the self-healing behavior that we're introducing here.