github/flagger
1
0
Fork 0
mirror of https://github.com/fluxcd/flagger.git synced 2026-02-23 14:24:09 +00:00
Code Issues Packages Projects Releases Wiki Activity
238 Commits 9 Branches 107 Tags
0.4.0
Commit Graph

76 Commits

Author SHA1 Message Date
stefanprodan
60f6b05397 Refactor scheduler tests 2019-01-18 11:14:27 +02:00
stefanprodan
6d8a7343b7 Add tests for analysis restart and canary promotion 2019-01-18 11:05:40 +02:00
stefanprodan
aff8b117d4 Restart validation if revision changes during analysis 2019-01-17 15:13:59 +02:00
stefanprodan
1d31b5ed90 Add canary name and namespace to controller logs 
- zap key-value: canary=name.namespace
 2019-01-17 13:58:10 +02:00
stefanprodan
1ef310f00d Add traffic weight to canary status 
- show current weight on kubectl get canaries and kubectl get all
 2019-01-16 16:29:59 +02:00
stefanprodan
acdd2c46d5 Refactor Canary status 
- add status phases (Initialized, Progressing, Succeeded, Failed)
- rename status revision to LastAppliedSpec
 2019-01-16 15:06:38 +02:00
stefanprodan
9872e6bc16 Skip readiness checks if canary analysis finished 2019-01-16 13:18:53 +02:00
stefanprodan
10c2bdec86 Use deep copy when updating the virtual service routes 2019-01-16 13:13:07 +02:00
stefanprodan
4bf3b70048 Use CRD UpdateStatus for Canary status updated 
- requires Kubernetes >=1.11
 2019-01-16 01:00:39 +02:00
stefanprodan
9232c8647a Check if multiple canaries have the same target 
- log an error on target duplication ref #13
 2019-01-15 21:43:05 +02:00
stefanprodan
23e8c7d616 Fix for when canary name is different to the target name 
- use target name consistent at bootstrap
 2019-01-15 21:18:46 +02:00
Stefan Prodan
42607fbd64 Merge pull request #26 from carlossg/service-name 
Fix VirtualService routes
 2019-01-15 19:38:38 +01:00
stefanprodan
28781a5f02 Use deep copy when updating the deployment object 
- fix canary status update logs
 2019-01-15 20:37:14 +02:00
Carlos Sanchez
5e880d3942 Wrong VirtualService routes 
If deployment name is different from canary name
the virtual service routes are created with canary name
but the services are created with deployment name

Note that canary name should match deployment name
 2019-01-15 18:44:50 +01:00
Stefan Prodan
3bff2c339b Merge pull request #20 from stefanprodan/scheduler 
Add canary analysis schedule interval to CRD
 2019-01-11 19:06:17 +01:00
Stefan Prodan
b035c1e7fb Merge pull request #25 from carlossg/virtualservice-naming 
Tries to create VirtualService that already exists
 2019-01-11 18:03:57 +01:00
Carlos Sanchez
7ae0d49e80 Tries to create VirtualService that already exists 
When canary name is different than deployment name

VirtualService croc-hunter-jenkinsx.jx-staging create error virtualservices.networking.istio.io "croc-hunter-jenkinsx" already exists
 2019-01-11 17:47:52 +01:00
Stefan Prodan
07f66e849d Merge branch 'master' into scheduler 2019-01-11 15:07:03 +01:00
stefanprodan
83118faeb3 Fix autoscalerRef tests 2019-01-11 13:51:44 +02:00
stefanprodan
aa2c28c733 Make autoscalerRef optional 
- use anyOf as a workaround for the openAPI object validation not accepting empty values
- fix #23
 2019-01-11 13:42:32 +02:00
Carlos Sanchez
c1bde57c17 Fix bad error message 
"controller/scheduler.go:217","msg":"deployment . update error Canary.flagger.app \"jx-staging-croc-hunter-jenkinsx\" is invalid: []: Invalid value: map[string]interface {}{\"metadata\":map[string]interface {}{\"name\":\"jx-staging-croc-hunter-jenkinsx\", \"namespace\":\"jx-staging\", \"selfLink\":\"/apis/flagger.app/v1alpha2/namespaces/jx-staging/canaries/jx-staging-croc-hunter-jenkinsx\", \"uid\":\"b248877e-1406-11e9-bf64-42010a8000c6\", \"resourceVersion\":\"30650895\", \"generation\":1, \"creationTimestamp\":\"2019-01-09T12:04:20Z\"}, \"spec\":map[string]interface {}{\"canaryAnalysis\":map[string]interface {}{\"threshold\":5, \"maxWeight\":50, \"stepWeight\":10, \"metrics\":[]interface {}{map[string]interface {}{\"name\":\"istio_requests_total\", \"interval\":\"1m\", \"threshold\":99}, map[string]interface {}{\"name\":\"istio_request_duration_seconds_bucket\", \"interval\":\"30s\"istio-system/flagger-b486d78c8-fkmbr[flagger]: {"level":"info","ts":"2019-01-09T12:14:05.158Z","caller":"controller/deployer.go:228","msg":"Scaling down jx-staging-croc-hunter-jenkinsx.jx-staging"}
 2019-01-09 13:17:17 +01:00
stefanprodan
882b4b2d23 Update the control loop interval flag description 2019-01-08 13:15:10 +02:00
stefanprodan
bec96356ec Bump CRD version to v1alpha3 
- new field canaryAnalysis.interval
 2019-01-07 01:03:31 +02:00
stefanprodan
b5c648ea54 Bump version to 0.3.0-beta.1 2019-01-07 00:30:09 +02:00
stefanprodan
e6e3e500be Schedule canary analysis based on interval 2019-01-07 00:26:01 +02:00
stefanprodan
b008abd4a7 Fix metrics server offline test 2018-12-27 12:43:43 +02:00
stefanprodan
cbf9e1011d Add tests for metrics server check 2018-12-27 12:42:12 +02:00
stefanprodan
6ec3d7a76f Format observer tests 2018-12-27 12:21:33 +02:00
stefanprodan
ab52752d57 Add observer histogram test 2018-12-27 12:16:10 +02:00
stefanprodan
df3951a7ef Add observer tests 2018-12-27 12:15:16 +02:00
stefanprodan
722d36a8cc Add webhook tests 2018-12-26 17:58:35 +02:00
stefanprodan
e86c02d600 Implement canary external check 
- do a HTTP POST for each webhook registered in the canary analysis
- increment the failed checks counter if a webhook returns a non-2xx status code and log the error and the response body if exists
 2018-12-26 14:41:35 +02:00
stefanprodan
0d96bedfee Add webhooks to Canary CRD v1alpha2 2018-12-26 13:42:36 +02:00
Stefan Prodan
03408683c0 Add details to Slack messages 
- attach canary analysis metadata to init/start messages
- add rollback reason to failed canary messages
 2018-12-06 12:51:02 +07:00
stefanprodan
f3904ea099 Use canary state constants in recorder 2018-11-27 17:34:48 +02:00
stefanprodan
8878f15806 Clean up isDeploymentReady 2018-11-27 17:11:35 +02:00
stefanprodan
5977ff9bae Add rollback test based on failed checks threshold 2018-11-27 17:00:13 +02:00
stefanprodan
c7e7785b06 Fix canary deployer is ready test 2018-11-27 15:55:04 +02:00
stefanprodan
4cb5ceb48b Rollback canary based on the deployment progress deadline check 
- determine if the canary deployment is stuck by checking if there is a minimum replicas unavailable condition and if the last update time exceeds the deadline
- set progress deadline default value to 60 seconds
 2018-11-27 15:44:15 +02:00
stefanprodan
042d3c1a5b Set ProgressDeadlineSeconds for primary deployment on init/promote 2018-11-27 12:10:14 +02:00
Stefan Prodan
769aff57cb Add Slack notifications for canary events 2018-11-25 11:44:45 +02:00
Stefan Prodan
c5930e6f70 Update deployment strategy on promotion 
- include spec strategy, min ready seconds and revision history limit to initialization and promotion
 2018-11-24 20:03:02 +02:00
Stefan Prodan
65bd77c88f Add last transition time to Canary CRD status 2018-11-24 15:48:35 +02:00
Stefan Prodan
580924e63b Record canary duration and total 
- add Prometheus metrics canary_duration_seconds and canary_total
 2018-10-29 21:44:43 +02:00
Stefan Prodan
1b2108001f Add Prometheus registry flag to recorder 
- fix tests
 2018-10-29 14:04:45 +02:00
Stefan Prodan
074e57aa12 Add recorder to revision tests 2018-10-29 13:43:54 +02:00
Stefan Prodan
e16dde809d Add recorder to mock controller 2018-10-29 13:34:28 +02:00
Stefan Prodan
4a8aa3b547 Add recorder component 
- records the canary analysis status and current weight as Prometheus metrics
 2018-10-29 11:25:36 +02:00
Stefan Prodan
344c7db968 Make golint happy and add codecov 2018-10-23 16:36:48 +03:00
Stefan Prodan
8e2f538e4c Add scheduler tests for initialization and revision 2018-10-22 20:14:09 +03:00