Nathan

Hmmm, I've been running into this identical issue on the webos client for a long time and I do not have http/2 enabled on either my nginx or the docker...

I just tested and confirmed this issue still exists on headscale_0.23.0-alpha2_linux_amd64.deb

Here's a (slightly scrubbed) netmap from the client while broken. I was using the exact ACL at the top of this issue with just the allow 0.0.0.0 removed. [netmap.txt](https://github.com/juanfont/headscale/files/13877293/netmap.txt)

My install is prod now unfortunately so I'll need to spin up a test environment. Hopefully can do that next week.

I think this data is sent to their logtail server which we can't reconfigure. https://tailscale.com/kb/1011/log-mesh-traffic

You're missing the port, and I think you need square brackets around your IP See: https://tailscale.com/kb/1018/acls/ > Tailscale IP | 100.101.102.103 | Only the device that owns the given Tailscale...

Looks like that's a client limitation. See https://github.com/tailscale/tailscale/issues/4727

The query I mentioned in #3649 seems to have just been replicated out and taken out a significant number of instances, including lemmy.world

Ok, I'm frustratingly close to figuring this out but I'm just completely stumped. First off, two triggers seem to be to blame: site_aggregates_comment_delete and person_aggregates_comment_count I'm focusing on the first...

The second bug was an easier fix: ``` CREATE OR REPLACE FUNCTION public.person_aggregates_comment_count() RETURNS trigger LANGUAGE plpgsql AS $$ begin IF (was_restored_or_created(TG_OP, OLD, NEW)) THEN update person_aggregates set comment_count =...