Uploaded image for project: 'Red Hat OpenShift Data Science'
  1. Red Hat OpenShift Data Science
  2. RHODS-1811

Prometheus alerts firing but not sending e-mails

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Not a Bug
    • Icon: Blocker Blocker
    • None
    • RHODS_1.1_GA
    • Monitoring
    • None
    • False
    • False
    • No
    • No
    • No
    • Yes
    • None

      Description of problem:

      If I fill up the jupyterhub notebook PVC over 90% the alert "user PVC is 90% used" is correctly fired but the e-mail notifying to the user is not properly send

       

      If you check the alertmanager logs you'll find this  error that could be the cause:

      level=error ts=2021-09-08T13:08:22.198Z caller=dispatch.go:309 component=dispatcher msg="Notify for alerts failed" num_alerts=1 err="user-notifications/email[0]: notify retry canceled after 16 attempts: establish connection to server: dial tcp: lookup smtp.corp.redhat.com on 172.30.0.10:53: no such host"

      Actual results:

      No e-mail alert sent to the address specified in redhat-ods-operator > Secrets > addon-managed-odh-parameters > notification-emai

      Expected results:

      E-mail with the alert sent to the address

       

      Reproducibility (Always/Intermittent/Only Once):

      Always

      Build Details: 

      RHODS 1.1.1-22

      Additional info:

      This happens in a OSD cluster where RHODS 1.1.1-22 has been installed with our install script (using jenkins)

       

      I'm running this code in the notebook to fill the pvc:

      import os
      
      # user notebook volumes are 2046640128 bytes
      sizeBelow90Percent = 1650000000
      sizeOver90Percent  = 1841976115
      size100Percent = 2046640128
      
      # Fill the volume over 90% to check that the alert is fired
      with open('large_file_over90', 'wb') as fout:
          fout.write(os.urandom(sizeOver90Percent))
      

       

       

       

              Unassigned Unassigned
              rhn-support-jgarciao Jorge Garcia Oncins
              Jorge Garcia Oncins Jorge Garcia Oncins
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

                Created:
                Updated:
                Resolved: