app_tokenAssociationsUsed metric is not consistent across nodes #17718
Labels
Bug
An error that causes the feature to behave differently than what was expected based on design.
Observability
Tickets pertaining to how observable the consensus node is (metrics, dashboards, etc.)
This bug used to be a part of #17062, but it's actually a separate problem, so I am filing a new ticket.
Grafana link (perf1 network): https://preproduction.grafana.hedera-ops.com/d/any-stat-prom/any-stat-prometheus?from=2025-01-30T06:08:29.000Z&to=2025-02-01T03:57:31.000Z&timezone=utc&var-Datasource=grafanacloud-prom&var-network=performance&var-name1=app_accountsUsed&var-name2=Reconnect_endsReconnectAsReceiver_created&var-name3=platform_secC2C&var-name4=platform_ev_per_syncS&var-name5=Reconnect_endsReconnectAsReceiver_created&var-metric1_type=$__all&var-metric2_type=$__all&var-metric3_type=$__all&var-metric4_type=$__all&var-metric5_type=$__all&orgId=1&var-name=app_tokenAssociationsUsed&var-metric_type=$__all
This dashboard shows
app_tokenAssociationsUsed
metric. My understanding this metric must be consistent (actually, equal) across all nodes, but it is clearly not. It may or may not be related to reconnects. Many of these nodes had experience a reconnect or multiple reconnects before the specified time period, but I don't know if this is the source of the problem.A screenshot is attached:
The text was updated successfully, but these errors were encountered: