Releases: opsani/servox
Releases · opsani/servox
v2.2.2 "genesis"
Changed
- Updated third party libraries
v2.2.1 <genesis>
Added
- Modify Opsani dev connector to work with Appd configuration
- Support for agent management sidecar
Fixed
- Fix servo crash on connection error
v2.2.0 "genesis"
Added:
- Appdynamics authentication
- Add helm chart
v2.1.1 "subtle rhapsody"
Added
- Servox preliminary support for optimization of StatefulSets 439
Changed
- README.md and checks documentation updated to reflect changes to CLI executions of checks 445
Fixed
- Kubernetes metrics server connector returning tuning memory request/limit from main workload 449
- Invalid operation error when comparing NaN values against configured SLO mimimums 450
Security
- httpx updated to version 0.23.0 446
v2.1.0 "subtle rhapsody"
Added
- Native MacOS support 422
- Container logs are now included in rejections caused by restarting pods 383
- Test coverage for container logs 394
- Added Black code formatter 398
- Added MyPy code scans and inline linting 402
no_tuning
configuratibility to allow external drivers to leverage Opsani Dev orchestration without being required to launch a Saturation or Tuning optimization 404- Combined CI coverage reports from Unit tests and Integration tests 425
Changed
- Pinned certain dependencies for improved stability:
- Response bodies are now included in the logging of unretryable HTTP responses 400
- Updated dependencies:
- Remedies from multiple failing checks are now applied in the same check iteration. Previous behavior was to apply a single remedy from the first failing check then rerun all checks, rinse and repeat until passing 411
- Fast Fail metric thresholds that are close to 0 will be skipped instead of counting as failures 412
- Updated kubernetes connector container
resources
logic to returnlimits
in place ofrequests
in cases whererequests
are not set 426
Fixed
- Improved CI test stability 388
- Resolved edge case where timeouts would run indefinitely if set to a duration of 0 392
- Resolved error in connector logic that could have caused the wrong target resource to be referenced as the target or failed to locate the target resource altogether 422
Removed
v2.0.0 "electric orchestra"
Added
- Kubernetes metrics connector for reporting Pod metrics from kubernetes metrics server #369
- Ability to upload logs and configuration for tech support #373
- Support for describing and adjusting kubernetes container environment variables #375
Changed
- Updated project to python 3.9.7 #367
- No longer using custom kubernetes_asyncio package #368
- The Opsani Dev connector attaches the Kubernetes Metrics connector if we're able to list pod metrics #379
Fixed
- Relaxed python version dependency requirements #372
v0.11.1 "preposterous ports"
v0.11.0 "preposterous ports"
Added
- Prometheus connector fast-fail support; short-circuit long running measurements when SLO violations are detected #330
- Kubernetes connector Argo Rollouts WorkloadRef support #337
- Opsani Dev connector support for configuration of image and tag used for envoy sidecar injection #341
- Kubernetes connector support for injecting static environment variables into the tuning Pod #343
Changed
- Kubernetes Deployment optimizations (Saturation Mode)
destroy
error behavior changed toshutdown
; the Deployment
resource is no longer destroyed and is scaled to zero replicas instead #317 - Kubernetes adjustment values no longer raise validation errors when outside the configured range (eg. falling back to
initially observed/baseline values) #279
Fixed
v0.10.7 "baseless allegation"
Changed
- Updated opsani Dev connector metrics to remove confusing _avg metric name #33
1
v0.10.6 "baseless allegation"
Changed
- Update forked kubernetes_asyncio dependency to use PyPI package source instead of github #315