Remember the good old days when monitoring a Drupal site was as simple as waiting for a panicked call from a client or end-user screaming "THE SITE IS DOWN!"? So do we. Fortunately or unfortunately, the rise of the cloud and contemporary DevOps practices have made monitoring infrastructure, and especially resource-intensive applications like Drupal, a complex science. Just ask the dozens of (totally awesome and good-looking) vendors offering SaaS-based monitoring for Drupal sites!
But is it really all that complicated? This talk will break down the jargon and provide a do-it-yourself roadmap for building a metrics- and logs-aggregation toolchain for Drupal infrastructure along with visualization and analytics tools, using off-the-shelf open-source software such as Elasticsearch, Logstash and Kibana. You'll learn cool lessons like:
- Building a rock-solid monitoring platform that keeps up with your evolving cloud architecture
- What metrics are most valuable in diagnosing a "sick" (as in ailing, not as in awesome) Drupal instance
- How monitoring and containers can play nicely together
- Using machine-learning algorithms to predict performance impairments before they happen
- Harnessing cloud scaling tools to make Drupal infrastructure self-healing
We can't promise to make that "URGENT!!!" 3am pager message a thing of the past, but we can hopefully show how to make it a lot more meaningful--and for free.