Uploaded image for project: 'OpenShift Logging'
  1. OpenShift Logging
  2. LOG-3856

spec.logStore.type == lokistack in clusterlogging/instance is the only way to enable the LokiStack Console Plugin

XMLWordPrintable

    • 2
    • False
    • None
    • False
    • NEW
    • Administer, API
    • NEW
    • Feature
    • Log Storage - Sprint 242

      Based on OBSDA-295, for transition purpose, it's possible to to run OpenShift Container Platform 4 - Cluster Logging with elasticsearch managed via clusterlogging/instance and have LokiStack run and managed in parallel via lokistack/instance. The only thing clusterlogging/instance will do when using LokiStack is to manage the default collector.

      This is all fine, but now it was found that the necessary LokiStack Console plugin is only created when spec.logStore.type == lokistack is set in clusterlogging/instance which is a hard blocker for the proposed solution. Given that it's possible to have both stacks in parallel and use ClusterLogForwarder to duplicate logs in both stacks, it should be possible to configure the LokiStack Console Plugin differently as spec.logStore.type == lokistack can not be set in this scenario.

      While as a work-around it would be possible to create the required Console Plugin manually, this is not desired as it would require manual updating, etc.

      Therefore having a way to enable the LokiStack Console plugin without spec.logStore.type == lokistack is a hard requirement respectively a blocker for the proposed solution.

            ptsiraki@redhat.com Periklis Tsirakidis
            rhn-support-sreber Simon Reber
            Anping Li Anping Li
            Votes:
            1 Vote for this issue
            Watchers:
            5 Start watching this issue

              Created:
              Updated:
              Resolved: