|Status:||In development, currently a proof of concept|
Goal is to easily generate metrics from your (primarily web) apps and write them somewhere useful.
Eventually it might take over some form of log handling too. The intent here would be to replace use of logging as random debug and performance aids and instead emit more structured information, or at least better correlate log messages with events.
How it works
For a simple WSGI app you install the middleware, which starts tracking "contexts". Each context is tied to a particular thread and stack (at the moment it assumes a 1:1 stack to thread correlation, for async frameworks this needs to be updated to handle switching of contexts).