At around 21:30 UTC, we received reports that shared runners were no longer processing jobs. Upon investigation we discovered that Digital Ocean, our provider for shared runners, was experiencing a [droplet creation issue in NYC1](http://status.digitalocean.com/incidents/ylv2nl83fb68). Since our runner managers are hosted in NYC1, we stopped processing jobs. To mitigate this issue, I unpaused the backup runners we have at GCE and they processed all of the pending jobs. At 22:32 UTC, Digital Ocean resolved the issue and I paused the GCE runners once more. cc/ @markpundsack @eReGeBe @pcarranza