Uploaded image for project: 'Red Hat Fuse'
  1. Red Hat Fuse
  2. ENTESB-12638

Fuse 7.5 on OpenShift 3.11: Prometheus collects no metrics

    Details

    • Type: Bug
    • Status: Closed
    • Priority: Major
    • Resolution: Cannot Reproduce
    • Affects Version/s: fuse-7.5-GA
    • Fix Version/s: fuse-7.7-GA
    • Component/s: Monitoring
    • Labels:
    • Steps to Reproduce:
      Hide

      1. Install Fuse 7.5 OpenShift templates on OpenShift 3.11, as described in the Red Hat documentation [1], section 2.5
      2. Obtain any sample application for Camel under Spring Boot, designed to be deployed on OpenShift. I used one from the automated builder at developers.redhat.com/launch (on the basis that, being an official Red Hat sample, it is likely to be correct)
      3. 'oc login' as an unprivileged user
      3. Deploy the application using 'mvn fabric8:deploy'. This creates, in my case, the deployment config 'newapp-fuse'
      4. Ensure that an application pod is running, and the application is doing whatever it should do
      5. Install Prometheus operator as described in the documentation [2]
      5a. oc login as system:admin
      5b. Install customer resource defs

      $ oc create -f  https://raw.githubusercontent.com/jboss-fuse/application-templates/master/fuse-prometheus-crd.yml
      

      5c. Install the Promethus operator

      $ oc process -f https://raw.githubusercontent.com/jboss-fuse/application-templates/master/fuse-prometheus-operator.yml -p NAMESPACE="myproject" | oc create -f -
      

      5d. Check the operator is running

      $ oc get pods|grep oper
      prometheus-operator-7567bb7b77-5c6hm   1/1       Running     0          20m
      

      5e. Set up the system monitor. In my example, the application's deployment configuration (and everything else) is called 'newapp-fuse'.

      $ oc process -f https://raw.githubusercontent.com/jboss-fuse/application-templates/master/fuse-servicemonitor.yml -p NAMESPACE=myproject -p FUSE_SERVICE_NAME=newapp-fuse|oc create -f -
      Error from server (AlreadyExists): services "newapp-fuse" already exists
      

      I do not know whether the 'services already exists' message is significant or not.

      6. Open the Prometheus console by, e.g., following the link from the 'prometheus' route in the OpenShift console

      7. Note that the drop-down list of exposed metrics is empty.

      [1] https://access.redhat.com/documentation/en-us/red_hat_fuse/7.5/html-single/fuse_on_openshift_guide/index
      [2] https://access.redhat.com/documentation/en-us/red_hat_fuse/7.3/html/managing_fuse/accessing_prometheus

      Show
      1. Install Fuse 7.5 OpenShift templates on OpenShift 3.11, as described in the Red Hat documentation [1] , section 2.5 2. Obtain any sample application for Camel under Spring Boot, designed to be deployed on OpenShift. I used one from the automated builder at developers.redhat.com/launch (on the basis that, being an official Red Hat sample, it is likely to be correct) 3. 'oc login' as an unprivileged user 3. Deploy the application using 'mvn fabric8:deploy'. This creates, in my case, the deployment config 'newapp-fuse' 4. Ensure that an application pod is running, and the application is doing whatever it should do 5. Install Prometheus operator as described in the documentation [2] 5a. oc login as system:admin 5b. Install customer resource defs $ oc create -f https: //raw.githubusercontent.com/jboss-fuse/application-templates/master/fuse-prometheus-crd.yml 5c. Install the Promethus operator $ oc process -f https: //raw.githubusercontent.com/jboss-fuse/application-templates/master/fuse-prometheus- operator .yml -p NAMESPACE= "myproject" | oc create -f - 5d. Check the operator is running $ oc get pods|grep oper prometheus- operator -7567bb7b77-5c6hm 1/1 Running 0 20m 5e. Set up the system monitor. In my example, the application's deployment configuration (and everything else) is called 'newapp-fuse'. $ oc process -f https: //raw.githubusercontent.com/jboss-fuse/application-templates/master/fuse-servicemonitor.yml -p NAMESPACE=myproject -p FUSE_SERVICE_NAME=newapp-fuse|oc create -f - Error from server (AlreadyExists): services "newapp-fuse" already exists I do not know whether the 'services already exists' message is significant or not. 6. Open the Prometheus console by, e.g., following the link from the 'prometheus' route in the OpenShift console 7. Note that the drop-down list of exposed metrics is empty. [1] https://access.redhat.com/documentation/en-us/red_hat_fuse/7.5/html-single/fuse_on_openshift_guide/index [2] https://access.redhat.com/documentation/en-us/red_hat_fuse/7.3/html/managing_fuse/accessing_prometheus
    • QE Test Coverage:
      +

      Description

      When installing Prometheus using the procedure in the Red Hat documentation, and deploying a Fuse 7.5 application built by Red Hat tools for OpenShift, Prometheus collects no metrics. The Prometheus port (default 9779) in the application pod shows a list of metrics, including Camel request metrics, and the Prometheus console is available. However, the Prometheus console shows no data, as if no applications were running.

        Gliffy Diagrams

          Attachments

            Activity

              People

              • Assignee:
                jondruse Jiri Ondrusek
                Reporter:
                kboone Kevin Boone
                Tester:
                Lukáš Löwinger
              • Votes:
                0 Vote for this issue
                Watchers:
                6 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: