Releases: nebula-orchestrator/manager
Releases · nebula-orchestrator/manager
2.7.0
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
2.5.0
2.4.0
2.3.0
2.2.0
2.1.0
2.0.1
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