1. Instrumenting autoscaling EC2 instances in Prometheus Using Service Discovery to overcome the laziness inherent to all sysadmins.
  2. Dealing with silent failures Your code isn't as good as you think it is, and that's OK.