Bob Weinand

With 1.5 the cluster agent default environment should be properly respected for traces and remote configuration.

Hey @iNviNho, Is there something in the internal state of `GlobalTracer::get()` which is ever-growing? I.e. observing `var_dump(GlobalTracer::get())`, do you observe some arrays growing or data structures getting more and more...

Hey @Julian-Louis, thanks for taking the initiative here! I think this approach may not be fully correct. Like, normalization is basically trying to recognize ids, but if you have routes...

Hmm @Julian-Louis, I think this code has another missing case: Now a route `/posts/{category}/{id}` matching `/posts/1/1` will be converted to `/posts/{category}/{category}`, if I read that code correctly? I've been looking...

I think the current solution is probably _viable_. It's not 100% correct (e.g. `/posts/posts` with a route `/posts/{name}` will be now `/{name}/posts`). I'm not sure if that's the only edge...

I think empty string handling is missing in Normalizer.php. I.e. it should also always match there with an empty string.

In fact I'm just seeing https://github.com/DataDog/dd-trace-php/pull/2909/commits/5f1f193bd1c5672272cff6a96952c631ac1adb57, is there a specific reason you reverted that? (it should just `return [$postKey => $postVal];` unconditionally, when the string is empty, too.)

Hey @junjihashimoto, thanks for the bug report, the main question here is: why is `_dd_writer_loop` started a second time, while there's already an active one? I haven't seen this happening...

@junjihashimoto Is it possible for you to reproduce this with valgrind or asan? Thanks!

Hey @brandonburkett, We've decided to keep full debug symbols of the rust code to enhance debuggability in case there are issues. However, this comes, as you noticed at the cost...