Add targetUtilization for percentile prediction #330
Conversation
2. Refactor percentile prediction
|
🎉 Successfully Build Images. Overview: https://finops.coding.net/public-artifacts/gocrane/crane/packages
|
| SampleInterval: sampleInterval, | ||
| MarginFraction: marginFraction, | ||
| TargetUtilization: targetUtilization, | ||
| Percentile: percentile, |
There was a problem hiding this comment.
Please make sure the go crane api is updated. now our structure is hard to extend, so maybe later we change the algorithm params to just a map to avoid the frequently updated to the crane api
| @@ -226,40 +193,7 @@ func (p *percentilePrediction) process(namer metricnaming.MetricNamer, config co | |||
| } | |||
| } | |||
|
|
|||
| estimator := NewPercentileEstimator(cfg.percentile) | |||
There was a problem hiding this comment.
Why remove this? this is used to do once task. it do not depends on the state signal cached in memory
| @@ -57,6 +58,46 @@ func generateSamplesFromWindow(value float64, start time.Time, end time.Time, st | |||
| return result | |||
| } | |||
|
|
|||
| func (p *percentilePrediction) getPredictedValuesFromSignals(queryExpr string, signals map[string]*aggregateSignal) []*common.TimeSeries { | |||
There was a problem hiding this comment.
we need abstract a common lib for algorithm functions
|
|
||
| targetUtilization, exists := props["resource.mem-target-utilization"] | ||
| if !exists { | ||
| targetUtilization = "1.0" |
There was a problem hiding this comment.
replicas.min-cpu-target-utilization: "30" replicas.max-cpu-target-utilization: "75" replicas.cpu-target-utilization: "50"
currently xxx-utilization is a 100 percent based float, better to align the same format for new configuration.
also please add it in deploy/craned/deployment.yaml and also the helm chart.
|
|
1. Add targetUtilization for percentile prediction 2. Refactor percentile prediction 3. Fix analytics reconcile loop issue Co-authored-by: yuyufei <yufeiyu@tencent.com>
What type of PR is this?
feat
What this PR does / why we need it:
Add targetUtilization for percentile prediction
Which issue(s) this PR fixes:
Fixes #
Special notes for your reviewer: