fix: sentinel readiness probe to fail when monitoring localhost

[ese]

Original PR: #558
Fixes #532 .

Changes proposed on the PR:

• Readiness check for sentinel shall fail when it doesn't monitor the right master, at least when we know that it monitors the wrong one which is 127.0.0.1
• There'll be dependency on grep and head existing on the redis image used for sentinels. I tested alpine and bullseye in the official redis images here and they worked as expected.

I think we should have customised readiness and liveness checks in the future as well but for now this should be enough.