fix reddis extension start func

[hilmarf]

Summary

Add connection tuning knobs to Redis storage config (timeouts, retry/backoff, ping timeout, pool timeout) with sensible defaults.
On startup, use the configured dial/read/write timeouts, ping for readiness, and retry with backoff, log attempts and return a clear failure when retries are exhausted.

Motivation

• The Redis extension could fail to start or silently ignore startup errors due to incorrect start logic.
• Users observed startup instability when the extension was enabled; this change makes startup behavior deterministic and failures visible.

What changed

• Corrected the Redis extension Start implementation to properly handle context, lifecycle, and returned errors.
• Improved error propagation so the collector fails fast on extension startup errors.
• Added/updated unit tests to cover the corrected startup behavior.
• Minor refactors to make the start flow clearer and more testable.

Testing

• Unit tests added to validate successful start and failure propagation.
• Manual verification: start the collector with the Redis extension enabled and confirm it fails on invalid configs and starts on valid configs.
• Run: go test ./... (or CI).

Compatibility / Breaking Changes

• No behavioral breaking changes for properly configured deployments.
• Deployments that relied on prior silent-failure behavior will now see startup errors (intended).

Documentation

• Add or update README/docs to note that the Redis extension will now propagate startup errors (recommended to validate config).