fix(logging): monitor configured service reachability

[kkovaacs]

Summary

• Add non-blocking reachability monitors for configured validator, sequencer, NTX builder, and remote prover endpoints.
• Warn on the first failed connection attempt, retry with capped exponential backoff, and log when the service becomes reachable.
• Sanitize endpoint URLs before logging them.
• Stop monitors cleanly during shutdown.

Partially resolves point 5 of #2330

Changelog

[[entry]]
scope       = "general"
impact      = "added"
description = "User-visible warnings for unreachable external service endpoints have been added, making potential mis-configuration easier to spot."

[sergerad]

Should this monitoring not be inherent in the connections that the sequencer has to the various components? Why is a separate monitor required?

[kkovaacs]

Could be, but then we wouldn't report unreachable dependencies for services that are lazy-initialized. (See the original report in issue #2330.)