Monthly Archives: August 2015

Good for production

My checklist for when a new service is ready for production.
I know, nothing new, no discovery here. But it happens that even across one team different engineers will put different weight to different items or have different lists. So here’s mine.

I won’t mention testing here, which belongs more to development phase.

  • monitoring (work metrics, resource metrics, events, alerts, dashboards, service health checks, end to end checks)
  • run books covering at least basic failures
  • logs shipped to central log, no writing to local disk
  • if for some reason log can’t be shipped to central log – log rotation, yes for Docker too
  • scaling runbooks or even better – scaling automation
  • performance testing scenario
  • backup scenario
  • chaos monkey scenarios

What’s yours? What would you add?