Bad news friends, last night (EU time) our API had a 44 minutes of downtime. Our apologies.
This outage was unfortunate as it was the worst in our 3 year history and follows on a run of 161 days of uptime, and I admit i was confident we’d finish out the year with 51 weeks of 100% uptime. We typically launch new code (as opposed to geodata which is updating fairly continually) a few times a day, and have an extensive set of automated tests that have to pass before anything can go live. Unfortunately this code performed great when run normally, but didn’t work well at high load.
Lesson learned, we’re now expanding our test suite to better replicate real world conditions.
Our apologies to everyone who was effected by the outage. Please feel free to contact us if you have any concerns.