[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: horizontal autoscaler does not get cpu utilization




----- Original Message -----
> From: "Den Cowboy" <dencowboy hotmail com>
> To: users lists openshift redhat com
> Sent: Wednesday, February 24, 2016 9:35:34 AM
> Subject: RE: horizontal autoscaler does not get cpu utilization
> 
> I don't know if this is maybe the issue?
> In my browser https://hawkular-metrics.xx.xx.com/hawkular/metrics/status
> {"MetricsService":"STARTED","Implementation-Version":"0.13.0-SNAPSHOT","Built-From-Git-SHA1":"7dee24acfcfb3beac356e2c4d83b7b1704ebf82x"}
> curl on my master or nodes:
> curl -X GET https://hawkular-metrics.xx.xx.com/hawkular/metrics/status
> curl: (6) Could not resolve host: hawkular-metrics.xx.xx.com; Name or service
> not known
> 
> I'm just describing the IP of the node where my router is in my local
> /etc/hosts
> like this: xx.xx.xx.xx hawkular-metrics.xx.xx.com

The router configuration is not used for the HPA and so not being able to resolve the hostname from within the node or container should not be an issue.

What the HPA does use is the API proxy.

You can check if Heapster is accessible via the API proxy through the following command:

curl -H "Authorization: Bearer XXXXXXXXXXXXXXXXX" \
       -X GET https://${KUBERNETES_MASTER}/api/v1/proxy/namespaces/openshift-infra/services/https:heapster:/api/v1/model/

Are there any other errors in the OpenShift logs [not the container logs for Hawkular-Metrics, Cassandra or Heapster, those appear to be working since you can see metrics in the browser]

> 
> 
> From: dencowboy hotmail com
> To: users lists openshift redhat com
> Subject: horizontal autoscaler does not get cpu utilization
> Date: Wed, 24 Feb 2016 13:56:24 +0000
> 
> I'm on Origin 1.1.3
> I've confgured the cluster-metrics (its in the openshift-infra project!). I'm
> able to see all the metrics (memory & cpu) on my metrics-tab.
> Now I try to create a simple autoscaler:
> oc autoscale dc/test --min 2 --max 15 --cpu-percent=70
> 
> I've edited the dc of my container so now it's using resources requests and
> limits.
> In my webconsole I see the 2 cirkles and the MiB and millicores used.
> 
> But
> oc get hpa
> NAME REFERENCE TARGET CURRENT MINPODS MAXPODS AGE
> test DeploymentConfig/test/scale 70% <waiting> 2 15 21m
> 
> Target CPU utilization: 70%
> Current CPU utilization: <not available>
> 
> I read it was normal that the current state was waiting in the beginning. But
> it's already on <waiting> for 21 minutes.
> How can I check what's going wrong?
> 
> The logs of the pod where I've created the autoscaler doesn't show anything
> different than normal:
> AH00558: httpd: Could not reliably determine the server's fully qualified
> domain name, using 10.1.1.14. Set the 'ServerName' directive globally to
> suppress this message
> 
> The logs of my metrics (but they seem to work fine):
> casandra: INFO 12:59:51 Initializing hawkular_metrics.leases
> 
> heapster:
> I0224 13:28:35.108008 1 manager.go:152] ScrapeMetrics: time: 107.715184ms
> size: 160
> I0224 13:29:05.000203 1 manager.go:79] Scraping metrics start: 2016-02-24
> 13:28:30 +0000 UTC, end: 2016-02-24 13:29:00 +0000 UTC
> I0224 13:29:05.118256 1 manager.go:152] ScrapeMetrics: time: 117.97843ms
> size: 160
> I0224 13:29:35.000209 1 manager.go:79] Scraping metrics start: 2016-02-24
> 13:29:00 +0000 UTC, end: 2016-02-24 13:29:30 +0000 UTC
> I0224 13:29:35.106305 1 manager.go:152] ScrapeMetrics: time: 106.009207ms
> size: 160
> I0224 13:30:05.000266 1 manager.go:79] Scraping metrics start: 2016-02-24
> 13:29:30 +0000 UTC, end: 2016-02-24 13:30:00 +0000 UTC
> I0224 13:30:05.113913 1 manager.go:152] ScrapeMetrics: time: 113.477259ms
> size: 160
> 
> metrics deployer:
> ++ '[' -n '' ']'
> Deploying the Heapster component
> ++ oc process hawkular-heapster -v
> IMAGE_PREFIX=docker.io/openshift/origin-,IMAGE_VERSION=latest,MASTER_URL=https://kubernetes.default.svc:443
> ++ oc create -f -
> serviceaccount "heapster" created
> service "heapster" created
> replicationcontroller "heapster" created
> + echo 'Success!'
> Success!
> 
> Hawkular-metrics:
> [0m[0m12:59:48,619 INFO [org.jboss.as] (Controller Boot Thread) WFLYSRV0051:
> Admin console listening on http://127.0.0.1:9990
> [0m[0m12:59:48,620 INFO [org.jboss.as] (Controller Boot Thread) WFLYSRV0025:
> WildFly Full 10.0.0.Final (WildFly Core 2.0.10.Final) started in 16058ms -
> Started 374 of 664 services (385 services are lazy, passive or on-demand)
> [0m[0m12:59:52,446 INFO
> [org.hawkular.metrics.core.service.MetricsServiceImpl]
> (metricsservice-lifecycle-thread) HAWKMETRICS100001: Using a key space of
> 'hawkular_metrics'
> [0m[0m12:59:52,561 INFO
> [org.hawkular.metrics.api.jaxrs.MetricsServiceLifecycle]
> (metricsservice-lifecycle-thread) HAWKMETRICS200005: Metrics service started
> 
> _______________________________________________ users mailing list
> users lists openshift redhat com
> http://lists.openshift.redhat.com/openshiftmm/listinfo/users
> 
> _______________________________________________
> users mailing list
> users lists openshift redhat com
> http://lists.openshift.redhat.com/openshiftmm/listinfo/users
> 


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]