Builder chunk batches locking (mainly to fix nodes_monitor working on same chunks)

[jeniawhite]

For FE PRs use the template

Purpose

• Fixes nodes_monitor builders that collide and work on same chunks

Checklist

• Code:
• User Experience: Taken a moment to think the effects on our user
• Failure handling and Comments on non trivial sections:
• Cross Platform: Supporting Win 7/8/10 Server 12, Linux, Mac
• Testing:
• Unit/System Tests: Added tests ... / Already covered by existing tests ...
• Tested with: npm test
• Supportability:
• Diagnostics info, Phone Home info, ActivityLog events, External Syslog:
• Upgrade:
• Mongo Schema Upgrade:
• Agents changes, Server Platform changes and New packages added to package.json:

Technical Debt Created

• Opened Move nodes_monitor builders to scrubber and improve keys_lock timing logic #2499

Fixed Issues References

• Fixes No (or wrong) rebuilds from cloud pool during migration #2116

[guymguym]

@jeniawhite Thought about waited_to_run again it's not the right solution (my bad).
The reason it is wrong to consider loading the chunks only if the lock waited is that there could be many cases where the race happens even without the lock wait, just because of the asynchronous nature of the database connection. Lets discuss.