Uploaded image for project: 'OpenShift Service Mesh'
  1. OpenShift Service Mesh
  2. OSSM-12035

Traffic graph caching and background compute

XMLWordPrintable

    • Icon: Epic Epic
    • Resolution: Done
    • Icon: Undefined Undefined
    • None
    • None
    • Kiali
    • None
    • Traffic graph caching and background compute
    • False
    • Hide

      None

      Show
      None
    • False
    • Release Notes
    • In Progress
    • Hide
      Traffic Graph Caching

      Kiali v2.21 introduces traffic graph caching. Enabled by default. When a user navigates to the Traffic Graph and renders the initial graph, Kiali will start a background job to regenerate the graph periodically, based on the refresh interval set in the UI. The background job will cache the resulting graph and return it on subsequent UI requests. This can greatly improve re-render times, especially for larger graphs. Note that the initial graph render time will be unchanged. It is still recommended to use “Manual” refresh when working with large meshes, in order to fully define the desired graph before performing the initial request. Any fundamental change to the graph definition will invalidate the cache and restart a new refresh job. Users can navigate away and then back to the traffic graph, and resume with the latest cached graph, within the timeout period (10m by default).

      Backend resource utilization may be affected, although is not anticipated to change significantly. The caching can be disabled in the Kiali config via:

      spec.kiali_internal.graph_cache.enabled: false


      Show
      Traffic Graph Caching Kiali v2.21 introduces traffic graph caching. Enabled by default. When a user navigates to the Traffic Graph and renders the initial graph, Kiali will start a background job to regenerate the graph periodically, based on the refresh interval set in the UI. The background job will cache the resulting graph and return it on subsequent UI requests. This can greatly improve re-render times, especially for larger graphs. Note that the initial graph render time will be unchanged. It is still recommended to use “Manual” refresh when working with large meshes, in order to fully define the desired graph before performing the initial request. Any fundamental change to the graph definition will invalidate the cache and restart a new refresh job. Users can navigate away and then back to the traffic graph, and resume with the latest cached graph, within the timeout period (10m by default). Backend resource utilization may be affected, although is not anticipated to change significantly. The caching can be disabled in the Kiali config via: spec.kiali_internal.graph_cache.enabled: false
    • Feature

      As part of our focus on Scalability, this focuses on the UX of the Kiali Traffic Graph. Once the initial graph is established, we want to compute updates to the graph on the server-side, caching the results. This should make subsequent client requests very fast [for the same graph, just advancing the time period]. Currently, graphs are generated at request time, potentially making the client wait a substantial amount of time.

      Note that this is only for namespace graphs, not node graphs (which includes graphs on the detail pages). It also has no effect on graph replay, it's only for advancing the current time period.

      This work can be tracked upstream via https://github.com/kiali/kiali/issues/8871

              rhn-engineering-jshaughn Jay Shaughnessy
              rhn-engineering-jshaughn Jay Shaughnessy
              Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

                Created:
                Updated:
                Resolved: