Graceful monitoring and alerting of multiple replica-healthchecks per namespace

Considering deploying an additional replica-healthcheck for the nodes consisting op-rbuilders, we'd have to ensure that our monitoring dashboards and alerts related to replica-healthcheck don't break.

They might break because a bunch of metrics tend to be aggregated on namespace or/and stack_optimism_io_network, both of which would be same for the rbuilder replica-healthcheck.

This issue tracks tackling that concerns and turning our dashboards and alerts accordingly.