-
-
Notifications
You must be signed in to change notification settings - Fork 4.8k
Description
Hello,
I was curious how interested you are in adding tracing with the OTEL libraries. I know we already have tracing thanks to PHP’s xdebug and I have already used it and it does the job. But the experience could be so much better.
At work we use the Golang OTEL library and, although I wasn’t convinced in the beginning, I must admit having traces flowing automatically from our apps to our Grafana instance (using Tempo) and having logs, metrics and traces all linked together is really, really great.
I checked the PHP OTEL library and they are even more mature than the Go ones. Also, it looks like we can have automatic instrumentation like xdebug does.
I did not try implementing anything yet and instead of dumping a PR out of the blue, I wanted to know if you were interested in having a PR adding tracing using the PHP OTEL library. If so, I’ve got no timeline guarantees but I’ll start work on this.
Btw, this all came from me wanting to squeeze out more performance from the WebDAV Nextcloud implementation.
EDIT: looking at some of the bugs and issues raised here and in the forum, I’m convinced a lot of these could be diagnosed thanks to traces.
EDIT2: also I’m maybe not seeing the full picture here. How do ppl usually track down performance issues in Nextcloud? So there’s xdebug, some amount of monitoring metrics, some logs, then more monitoring on external things used by Nextcloud like disk I/O, CPU and memory usage, etc. Apart from xdebug, I don’t see these as telling you exactly where the issue lies but instead as helpers to pinpoint where to look. Am I missing something?
Metadata
Metadata
Assignees
Labels
Type
Projects
Status