AWS Outage on 10/20

[desmos-api-announce]

Hi all,

In the last few weeks, we’ve been updating our incident response and mitigation plans and had the chance to put them to the test sooner than expected with yesterday’s substantial AWS outage, which impacted our database hosting service (Heroku). Those of you who rely on our website for creating graphs might have noticed unavailability of saving between roughly 12:10pm ET and 3:50pm ET.

We began monitoring the situation before it affected us, and are pleased to report that we maintained virtually uninterrupted service to our hosted API. We successfully served 99.96% of API requests yesterday, with three brief spikes of errors due to automated failovers as Heroku fought to restore service.

Going forward, we’ve created a status dashboard at status.desmos.com. We recommend subscribing to updates on that website. That will be our primary way of communicating about any incidents going forward, both for our API and our website. For detailed historical uptime tracking, we will also continue to maintain our Nodeping API uptime tracker. 

Again, we’re so grateful for your ongoing partnership,

The Desmos Studio Team