We just started looking into Datadog, but from what we see, it's like New Relic meets Loggly. It's really easy to plugin different services (like the one on this list) and get detailed analysis of what is happening on your servers and services. It makes tracking down sparse and difficult to understand problems possible.
Monitoring day-to-day operations of multiple high-performance computing assets distributed across several networks. Monitoring vendor provided data and setting up alerts when things do not show up on time.