metrics docs added and few changes to installation doc

INSTALLATION.md

@@ -69,7 +69,7 @@ their default values.
 
 | Parameter | Type | Default | Description |
 |-----------|------|---------|-------------|
-| `ottoscalr.config.metricsScraper.prometheusUrl` | string | `""` | URL where prometheus for the kubernetes cluster is running.   |
+| `ottoscalr.config.metricsScraper.prometheusUrl` | string | `""` | URL where prometheus for the kubernetes cluster is running. Fetching metrics from a single or multiple prometheus instance(give comma separated urls) is supported. Metrics from multple prometheus instances will be aggregated. If you have 2 instances named `p8s1` and `p8s2`, it should be added like `"p8s1,p8s2"`  |
 | `ottoscalr.config.metricsScraper.queryTimeoutSec` | int | `300` | Time in seconds within which the response for any query should be served by the prometheus  |
 | `ottoscalr.config.metricsScraper.querySplitIntervalHr` | int | `24` | The shortest period in hour for which data will be fetched from prometheus |
 | `ottoscalr.config.policyRecommendationController.maxConcurrentReconciles` | int | `1` | Maximum number of concurrent Reconciles which can be run. |

charts/ottoscalr/envconfigs/ottoscalr_config.yaml

@@ -6,19 +6,19 @@ enableLeaderElection: false
 leaderElectionID: "85d48caf.fcp.ottoscalr.io"
 metricsScraper:
   prometheusUrl: {{ .Values.ottoscalr.config.metricsScraper.prometheusUrl }}
-  queryTimeoutSec: {{ .Values.ottoscalr.config.metricsScraper.queryTimeoutSec | default "30" }}
+  queryTimeoutSec: {{ .Values.ottoscalr.config.metricsScraper.queryTimeoutSec | default "300" }}
   querySplitIntervalHr: {{ .Values.ottoscalr.config.metricsScraper.querySplitIntervalHr | default "24" }}
 breachMonitor:
   pollingIntervalSec: {{ .Values.ottoscalr.config.breachMonitor.pollingIntervalSec | default "300" }}
   cpuRedLine: {{ .Values.ottoscalr.config.breachMonitor.cpuRedLine | default "0.75" }}
   stepSec: 30
   concurrentExecutions: {{ .Values.ottoscalr.config.breachMonitor.concurrentExecutions | default "50" }}
 periodicTrigger:
-  pollingIntervalMin: {{ .Values.ottoscalr.config.periodicTrigger.pollingIntervalMin | default "1440" }}
+  pollingIntervalMin: {{ .Values.ottoscalr.config.periodicTrigger.pollingIntervalMin | default "180" }}
 policyRecommendationController:
   maxConcurrentReconciles: {{ .Values.ottoscalr.config.policyRecommendationController.maxConcurrentReconciles | default "1" }}
   minRequiredReplicas: {{ .Values.ottoscalr.config.policyRecommendationController.minRequiredReplicas | default "3" }}
-  policyExpiryAge: {{ .Values.ottoscalr.config.policyRecommendationController.policyExpiryAge | default "48h" }}
+  policyExpiryAge: {{ .Values.ottoscalr.config.policyRecommendationController.policyExpiryAge | default "3h" }}
 policyRecommendationRegistrar:
   requeueDelayMs: 500
   excludedNamespaces: {{ .Values.ottoscalr.config.policyRecommendationRegistrar.excludedNamespaces | default "kube-system,monitoring,gatekeeper-system,webhook" }}
@@ -28,7 +28,7 @@ cpuUtilizationBasedRecommender:
   stepSec: 30
   minTarget: {{ .Values.ottoscalr.config.cpuUtilizationBasedRecommender.minTarget | default "5" }}
   maxTarget: {{ .Values.ottoscalr.config.cpuUtilizationBasedRecommender.maxTarget | default "60" }}
-  metricsPercentageThreshold: {{ .Values.ottoscalr.config.cpuUtilizationBasedRecommender.metricsPercentageThreshold | default "0" }}
+  metricsPercentageThreshold: {{ .Values.ottoscalr.config.cpuUtilizationBasedRecommender.metricsPercentageThreshold | default "25" }}
 metricIngestionTime: 15.0
 metricProbeTime: 15.0
 enableMetricsTransformer: {{ .Values.ottoscalr.config.enableMetricsTransformer | default false }}

charts/ottoscalr/values.yaml

@@ -84,8 +84,7 @@ ottoscalrConfigPath: "/etc/ottoscalr/config/ottoscalr_config.yaml"
 ottoscalr:
   config:
     metricsScraper:
-      prometheusUrl: http://prometheus-k8s-0.prometheus-operated.monitoring.svc.cluster.local:9090,http://prometheus-k8s-1.prometheus-operated.monitoring.svc.cluster.local:9090,http://prometheus-k8s-2.prometheus-operated.monitoring.svc.cluster.local:9090
-      queryTimeoutSec: 300
+      prometheusUrl: 
       querySplitIntervalHr: 24
     policyRecommendationController:
       maxConcurrentReconciles: 1