syncInterval update event fix for duplicate/ghost events
cpitstick-argo opened this issue · comments
The code that processes Kubernetes events regularly runs a synchronization that causes every single event to trigger an update in a cluster periodically. However, these updates contain no new information. They are effectively doubled up.
When this happens, eventrouter will mark duplicate events, and events that have already been processed will appear as "ghosts" or "echoes of the past." This can cause a lot of confusion for any monitoring dashboard (such as a Prometheus/Grafana dashboard) which relies on eventrouter.
Containership fixed this in their fork a few years ago: https://github.com/containership/eventrouter/commit/f62fe77a43bf06fd846acb73ee4d473186a8ef5b
Please fold in this fix (or otherwise fix it in a different way, whatever you deem appropriate).
For more information about what a monitoring dashboard looks like with these duplicate events, see this issue: litmuschaos/litmus#1472