Related to the outage that happened last night, we rebooted the Lemmy services but we're still trying to figure out the root cause, which seems to point to an out of memory issue in the logs. However it's not what we see in our monitoring console.
In the meantime, we will monitor the service more closely until we are confident the issue is resolved, and we will improve our tools to detect such a problem faster.
EDIT: Also happened at night on July 2nd, still trying to find the root cause..
As I suggested elsewhere, could you maybe setup a "status" community on another instance (e.g. sh.itjust.works, it's Canadian as well), so that people can go there to see updates about the potential outages?
I'm looking at making a custom CloudFlare error page that embeds the status page. At least we'll be able to put some communication there when something happens without people having to guess where to go.