The outage seemed to be limited only to SSL requests. Since all non-ssl requests to .heroku.com are forcefully redirected to ssl, all .heroku.com requests failed during the outage.
We were able to handle requests at our http://*.herokuapp.com endpoint, but all https endpoints were down during the outage.
Just resolved it seems. Hit us in the middle of a huge enterprise demo. We’re moving off Heroku as fast as we can, but we need to make the jump to containerization first and it’s hard to prioritize infra projects these crazy days.
Would be interested in hearing what heroku users are using for automated failover when heroku goes down.
Obviously can't live in the application code (b/c that's unreachable at that point), but is there an easy way to configure Cloudflare or something to switch to a static site hosted somewhere else?
Cloudflare itself will host a cached version (plus a little Cloudflare-injected banner at the top) in this situation, but not apparently not all of the time reliably. It has worked for us in other "520" error situations.
Seems up at the moment. I'm sure there will be some outages as data centers are running with reduced staff and some companies don't know how to do teleworking well.
It currently takes you to Heroku's homepage, which contains no information relevant to the post's title.