Implementing the data store for the metrics, the ability to collect them, and the ability to retrieve them (Cassandra, collectd, statsd, cyanite, graphite-api, Grafana, cyanite-proxy). Implementing a distributed monitoring system -- get notified if something breaks without relying on a centralized system.
Presenting language plugins for statsd to get developers excited about adding statistics to their code for easier development planning.
Make the JSON dashboards are part of a project's code, so when code is updated, the dashboards are updated as well!
Showing some of Change.org's use cases, along with some of the pitfalls we ran into along the way.
Making monitoring not suck!
Jeff Pierce has been working in systems administration and DevOps for over a decade, and has worked for companies such as Apple and Rackspace and consulted for Citigroup. He's been a part of enterprise organizations and small startups, and has experience scaling solutions for companies of all sizes.