Holger Weiß

> My 'load' is 1 client ;) And the OS is bored as well? Either way, seems you can reproduce the issue better than others, who run into it only...

Next time you run into the issue, could you check whether running the following call in an `ejabberdctl debug` shell fixes the issue: ```erlang fast_tls:clear_cache(). ``` (Including the trailing `.`....

The problem being related to `fast_tls`s caching was just a blind guess of mine. I suggested that test to make sure nothing else is reloaded, just to track things down.

I think such a draft should also support deletion of [HTTP-uploaded][1] resources when deleting a message that references an uploaded file. The problem with supporting this cleanly is that the...

The one problem I'm aware of is that while globally configured certificates do work initially for `ejabberd_stun` listeners, it falls apart once you renew the certificate(s). This is because `ejabberd_stun`...

> But force_node_config should enforce also things when they are requested from the DB AFAIK. I totally see how that behavior is probably desired in most cases, but I do...

I'm trying my best not to troll this issue, but I really don't grasp why you'd reinvent features such as federation and history sync (which are anything but trivial to...

> The DFN-CERT sent the warning below for our STUN server We got the same email :smile: Unfortunately, there's [no proper fix][1], as: 1. per the spec, servers reply to...

> Would be cool to add an endpoint for monitoring! Maybe return a number of active connections? Have you seen [`mod_stats_prometheus`][1] and the [metrics it returns][2]? [1]: https://eturnal.net/documentation/#mod_stats_prometheus [2]: https://github.com/processone/eturnal/commit/07d7843486f87d9b1cab1b6084b78dd1629d99de

> In terms of security is better to put a proxy server like Nginx in front of Prometheus port? The idea is to restrict which IPs can access that endpoint....