I'm happy to see more independent documentation of this tool. It has helped a project I'm working on a great deal in the past week or so, allowing us to diagnose several bottlenecks, including extra gc, in our applications.
Maybe Pusher could donate some of their thoughts to the doc on trace, as it's quite thin.
I appreciate you explicitly calling out when `trace` is not appropriate (and pointing out how `pprof` is a good complement): Too many people look for "silver bullet" tools.
Good writeup, I didn't even realise this existed, thanks!
I think I'm going to hook in a little webservice that allows you to trigger a 1 or 5 second trace in a running prod app so I can run this ad hoc and see what's what in my various services.
5 comments
[ 2.9 ms ] story [ 19.8 ms ] threadMaybe Pusher could donate some of their thoughts to the doc on trace, as it's quite thin.
[1]: https://github.com/golang/go/blob/master/src/cmd/trace/doc.g...
I think I'm going to hook in a little webservice that allows you to trigger a 1 or 5 second trace in a running prod app so I can run this ad hoc and see what's what in my various services.