Which use case/requirement will be addressed by the proposed feature?

Currently the Redis sink handles back pressure in cases of OOM in a Redis target. However the current error messages are defined as warning and do not convey that the error is handled and retried.

Implementation ideas (optional)

Improve error messages and change severity from Warning to Info in two cases:
1. When memory threshold is set and memory has reached the threshold

2. When memory threshold is not set and target Redis is OOM