Releases: AMEST/swarm-autoscaler
Releases · AMEST/swarm-autoscaler
Decrease Mode and customize rules
Fix when no running tasks in service
Full Changelog: v0.2.2...v0.2.3
Fix getting container id's only for running containers
Fix getting leader status
Fix getting leader status when managers count more then 1
Optimizations and Improvements
What's Changed
Optimizations
- Move autoscaler code to own class extends threading.Thread
- Added thread pool for run calculating service cpu utilization in own thread.
- Added thread pool for discovery
- Added caching for: resolve cluster ip and node count inside calculate method
- Added disable manual replicas mode for service
Fixes
- Settings names in ReadMe
- Logging levels for third party libs
- Own loggers inside classes
Full Changelog: v0.1.0...v0.2.0
Minimum working version of the application
Minimum working version of the Swarm Services Autoscaler
application
Implemented:
- Automatically increase and decrease the number of service instances based on the average cpu of service instances
- Limits for the minimum and maximum number of service instances
- Independent calculation of cpu utilization by all instances of the service (without being tied to external data sources (cadvisor / telegraf + prometneus)
- Configurable cpu utilization limit at which there is a decrease in the number of replicas or their increase
- Configurable check interval
- Small documentation in README.md
- Example docker-compose file for application deployment