2.7.0

• Added option to force update all params of cron_jobs API endpoint

2.6.2

• Updated multiple prerequsite packages
• New version of the underlaying Alpine container OS
• Added .env file support for the configuration changes
• Adding report_on_update_only option
• Added updated field to the optional reporting system & filtering by it
• Multiple smaller bugfixes & improvements

2.6.1

• support for multiple types of configuration files\envvars\cli-args via parse_it
• updated prereqs
• added codecov
• added automatic creation of changelog
• multiple smaller bugfixes & improvements

2.5.0

• Updated prereqs
• Cron jobs support added
• Refactored index creation
• Multiple smaller bugfixes & improvements

2.4.0

• making travis deploy on all branches
• adding user groups
• multiple smaller bugfixes & improvements
• added authorization logic
• jsonfyied replies

2.3.0

• Added Travis CI/CD based build & push to docker hub
• Added multi user auth support
• Added bearer token auth support
• multiple minor bugfixes & improvements.

2.2.0

• Added optional reporting system
• Multiple smaller bugfixes & improvements
• Deprecated backend MongoDB _id field being returned

2.1.0

• Moved to Python 3.7
• Upgraded prereqs
• Docker image based on alpine3.9
• Added more unit tests
• Added code of conduct
• Token auth option prep work
• Added sane defaults to app configurations
• Multiple smaller fixes & improvements

2.0.1

• version matching to the worker

2.0.0

This release is complete redesign of Nebula with a new architecture, due to the scale of the changes it also includes breaking changes as both the design & the workflow is now radically improved.

• moved the backend MongoDB to it's own class
• basic auth error are now caught
• restructured the config files location
• folder structure refactor
• shippable ARM build python version is now locked
• no more RabbitMQ needed anywhere in the system
• fixed return code on some errors not being truly boolen
• added retrying to all read only functions with sane defaults
• added TTL based memomized cache logic on the device_group /info endpoint with LRU max size limit
• moved API to v2
• added monotonic ID to all changes in the DB
• added protection against edge cases when trying to add an app that doesn't exist to a device_group
• added rolling restart to the app config
• remove roll endpoint
• changed default mongo max pool size to 25
• added pruning on all devices
• added pruning on a per device_group basis
• added device_group logic - each device group consists of a list of apps it's running and each device\worker is configured at the worker container start to be part of a single device_group
• manager can now run with no conf.json file needed
• multiple syntax & pep8 fixes
• updated documentation