You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Beacon node for some nodes may lag a lot behind the head slot. We should alert this to the node operator on the grafana Healthy panel.
🛠️ Proposed solution
Add another const to app/metrics.go:
// readyzBeaconNodeFarBehind indicates that readyz is returning 500s since the Beacon Node is too far behind the head slot.
readyzBeaconNodeFarBehind = 8
Add the following logic to app/monitoring.go:
if beaconNodeDown {}
else if beaconNodeSyncing {}
else if beaconNodeZeroPeers {}
# The following is a new addition
else if beaconNodeFarBehind { readyzGauge.Set(readyzBeaconNodeFarBehind) }
...
BN far behind is TRUE if beacon_node_sync_distance > 320 slots. This is a fair threshold as 320 slots correspond to 10 epochs which is around 60 minutes.
So, it should potentially follow this
if beacon_node_not_syncing AND beacon_node_sync_distance > 320 slots { "BN far behind" }
If the panel shows "BeaconNode far behind" means that the operators need to check their Beacon Node for potential syncing issues.
🧪 Tests
Tested by new automated unit tests
The text was updated successfully, but these errors were encountered:
Adds "BeaconNode far behind" metric value to `/readyz`. This would help node operators to monitor their beacon node health if the `Healthy` panel shows `BeaconNode far behind` for a longer time.
category: feature
ticket: #1967
Adds "BeaconNode far behind" metric value to `/readyz`. This would help node operators to monitor their beacon node health if the `Healthy` panel shows `BeaconNode far behind` for a longer time.
category: feature
ticket: #1967
🎯 Problem to be solved
Beacon node for some nodes may lag a lot behind the head slot. We should alert this to the node operator on the grafana
Healthy
panel.🛠️ Proposed solution
Add another const to
app/metrics.go
:Add the following logic to
app/monitoring.go
:BN far behind
is TRUE ifbeacon_node_sync_distance > 320 slots
. This is a fair threshold as 320 slots correspond to 10 epochs which is around 60 minutes.So, it should potentially follow this
If the panel shows "BeaconNode far behind" means that the operators need to check their Beacon Node for potential syncing issues.
🧪 Tests
The text was updated successfully, but these errors were encountered: