Release 2.0

Major Changes

• Python process metrics - The reporter can now collect and report metrics about the system resource usage by
  the Python process it is running in. It can collect data about the CPU times, memory, GC, threads, and IPC. This
  is disabled by default.
• Switch to Prometheus compatible metric names - The reporter and client take a new parameter sanitize_mode
  which sanitizes the metric names to either Prometheus or Apptuit naming convention. The default mode is Prometheus. If you are upgrading from a previous release, you need to set the sanitize_mode parameter to apptuit if you want to stick to the old naming convention (dot separated metric names), otherwise after upgrade all the old metrics will be reported using the Prometheus naming convention (underscore separated).
• The reporter will add a host tag to all the metrics. It can be overridden in the global tags or metric tags.
• Dropped Python 3.4 support (it reached its end of life)

Release V1.4.0

Changes

• Deprecated the old environment variable names for token and global tags and introduced APPTUIT_API_TOKEN and APPTUIT_TAGS as new names. Old variables names are deprecated but still supported. Support for them should be removed in future releases.
• Removed pandas from hard requirements. Not everyone uses the client to query the data or create dataframes. Therefore it has been made a soft dependency which needs to be installed by hand. Documentation reflects this change as well.
• Added handling of 413 status code from PUT API - now we inform the user of the payload size which caused the error and the number of points which were sent as part of that payload, so that they can adjust their payload sizes to avoid the error.
• Split the payload in batches of 50,000 in the reporter
• Added a User-Agent header in all the requests being made by the client. The header includes the client version, the Python version and version of the requests module.

Release 1.3.0

Changes

• Removed validation on tag values as Apptuit backend does not impose those restrictions anymore
• Add check for maximum number of tags in datapoints and timeseries
• Add timeout parameter to send and query APIs
• Document about restrictions on tag and metric names

Release 1.2.1

Changes

Minor release - updated documentation for the changes introduced in release v1.2.0

Release 1.2.0

Changes

Added new API to the client send_timeseries to send a list of TimeSeries objects, as opposed to sending
a list of DataPoint objects with the send API.

Release 1.1.0

Changes

The Apptuit client and reporter were sending the data to the Apptuit backend in synchronous mode, which could be slow in certain scenarios. This release makes the data reporting asynchronous. Beyond this no functional changes are intended in this release.

Release 1.0.0

New Features

• Added support for environment variables for global tags and token
• Added meta metrics reporting - metrics about the data being reported
• Added error handler support to the pyformance reporter