All of Code.org’s infrastructure runs on Amazon Web Services, and they mostly rely on New Relic, HoneyBadger and Pingdom in addition to manual scripts for monitoring. Before joining PagerDuty, however, there was just one ops person in place to triage problems. They didn’t have a clear triage process or a single system of record, and that made it hard to delegate responsibility for issue resolution. The personal responsibility the one person took for getting things fixed was heroic, but a man can only go so long without sleep. A few engineers on the team had used PagerDuty at a previous company, and they saw that it offered the solutions they were looking for. Partnering with PagerDuty allowed Code.org’s nimble team of engineers to switch to a more distributed model.

Join for free to read