fix: Add additional serving bundle checks

[samcm]

Some Checkpointz instances have experienced lag on the 'serving' checkpoint, which is the checkpoint that checkpointz is actually serving on the finalized state id.

This PR does the following in an attempt to resolve this:

• Add additional triggers for checking if we can serve a new bundle
• Add mutexes to the processes that fetch bundles from beacon nodes to eliminate any chance of a data race
• Add additional logging in case the issue still persists