Port CronChecker

[pbiggar]

• Add healthcheck
• Add service
• Add container
• Add telemetry
• Port code
• Port tests
• determine what confidence we have that this works and if new tests are needed
• determine how we're going to switch over to this

[pbiggar]

Cronchecker doesn't interact with user code, so this feels safe to deploy now.

• Deploy an instance of cronchecker with code commented out
• verify pod works
• Look at memory usage, and set appropriate memory limits for .NET and k8s
• check that events are being processed in honeycomb
• enable everything except the actual enqueuing and record exucution.
• Validate that the events we expect are happening
  • Reduce huge number of events to managable number
  • Debug missing events
  • Re-add db events with a processor
• Deploy to production
  • Re-deploy with the code enabled and container count set to zero
  • Briefly pause the old cron-deployment (set container count to zero) and enable the new cronchecker-deployment (set container count to 1)
  • Verify it works by having cron stuff running
  • If it continues to work, leave the old cron-deployment with 0 containers
• Parallelize so that there's no skew in the crons