Sorry for adding to the noise here with all these posts about downtime an ping services, but I thought I’d add that I’ve been getting all sorts of problems, across all of my glitch servers (including boosted ones). They seem to be randomly restarting (which is fine), but then erroring* during the installation after the restart. So I have to open each project, make an edit to the watch.json file to trigger a re-install, and then it starts up fine. I have had a bunch of down-time that I didn’t know about because of the ping service blocks.
The combination of doing platform maintenance and blocking the down-time monitors at the same time has meant that I’m constantly having to do my down-time checking “manually” each day.
I’d love to be able to pay for more boosted apps, and to have the ability to pay to allow down-time monitoring on my Glitch servers. Is there any chance that the firewall could let through ping-service traffic to paid users’ servers, or even just to boosted apps?
For now I think I’m going to have to build my own mini down-time service. Not a big deal, but definitely not ideal.
In any case, I understand that sometimes some sacrifices need to be made for the long-term health of the platform, so good luck to the team on their spam-fighting and system-upgrading!
Thanks,
Joe
*As well as “module missing” installation errors (which are fixed by triggering a re-install by editing watch.json), I’m also sometimes seeing errors like “/opt/watcher/app-types/node/install.sh: line 21: wait_for: No record of process 1747” after editing the watch.json in general (i.e. when I edit the watch.json for some other reason - not due to installation errors). When this happened yesterday the application eventually restarted like normal, but I’ve got one server that is just stuck on that error. Going to try deleting and un-deleting next.