Elastic-apm

Bugfixes

* Updated CLOUD_PROVIDER config to allow for new options defined in 289 (878)
* Fixed a bug in AWS metadata collection on docker containers in AWS Elastic Beanstalk (884)
* Fixed a bug in local variable serialization that could occur in multi-threaded applications (885)

Features

* Added graphql (graphene) support [850]
* Collect cloud provider metadata [826]
* Added `sanitize_field_names` config [860]

Bug fixes

* Fix header fetching for ASGI implementations [856]

Features

* capture number of affected rows for INSERT/UPDATE/DELETE SQL queries (614)
* Added instrumentation for AsyncElasticsearch (843)

Bug fixes

* Added error handling around frame processing in Django (837)
* Limit SQL queries in context data to 10000 characters (842)
* Omit the "sync" property on spans by default (854)
* Update the pid/ppid in transport metadata when they change (825)
* Added better error handling around closing the transport (838)
* Fixed an issue with the django admin command not detecting misconfigured server urls (855)

New features

* Added `SERVICE_NODE_NAME` config [592]
* Added instrumentation support for [`pymemcache`](https://github.com/pinterest/pymemcache) [#812]
* Added new configuration options `enabled` and `recording` [790]

Bug fixes

* Fixed an issue with metrics collection raising RuntimeErrors [802]
* Fixed an issue with getting extended destination data from pymongo [797]
* Fixed an issue with Tornado instrumentation when Tornado isn't the "main" framework [813]
* Fixed an issue with streaming requests being consumed by the agent even if capture_body is disabled [817]

Pending Deprecations

* Python 2.7 support will be dropped in version 6.0 of the agent.

Bug fixes
* Fixed an issue with Redis using unix domain sockets and destination information (766)

Bug fixes
* Fixed Starlette middleware when capturing request body while the content-type header is absent (763)