- Don't update the log factory, user must install it manually.
- Remove python2.7 support.
- Migrate packaging to poetry.
- Remove python2.6 support.
- Add trace_id and span_id in the python logging context.
- Fix django issue on missing trace.
- Fix issue for django on slow request logs. The reset of the stack must be done on every requests.
- Ensure we never raise an exception if we cannot collect a trace, avoid side effect to clients.
- Hotfix Pyramid slow query logger that does not cleanup the trace stack.
- Rewrite Pyramid binding.
- Add settings for pyramid to log slow queries only with a configurable time.
- Add a setting to list what library should be traced.
- Make socket timout configurable and change default timeout to 1 second.
- Make the scribe async/sync socket configurable
- Add a middleware/setting for Django to track only slow query
- Add a Trace context manager for more flexibility
- Add django support using a django middleware and app.
- Add psycopg2 cursor support to trace sql query
- Add an http client (synchronous) to push trace, client transport is configurable.
- update logging level to avoid errors logs for SQL query outside http context
- remove deprecated log.warn
- requests: fixup infinite recursion
- pyramid: fixup tests if zipkin not configured
- pyramid: fixup tweenview init
- pyramid: register trace in a tweenview
- advertise version in pyramid module
- do not throw exception when trying to format to thrift
- ensure zipkin does not raise when trace id is larger than expected
- do not throw warning on configuration mistakes
- fixup python2 support
- refactor pyramid plugin
- fix reporter
- Use thriftpy2
- Flask bindings
- xmlrpclib client bindings
- Filtered parameters in sqlalchemy binding
- Implement exponential backoff on connection
- Make the service name configurable for pyramid application
- Keep @trace usable when zipkin is not configured
- Initial version