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.