⌘ / Ctrl K

Troubleshooting

Start in Diagnostics and Data health. Each blocker below names the cause and the fix.

On this page

Pipeline blockers

Blocker	Meaning	Fix
`NO_INTEGRATION`	No billing source is connected	Add an integration in Settings
`INTEGRATION_FAILED`	The last billing run failed	Check the role or identity, then reset the integration
`NO_SNAPSHOTS`	No agent has reported	Check the CronJob, its log and the cluster's token
`SNAPSHOTS_STALE`	A cluster stopped reporting	Check the latest job's log in the `kubeon` namespace
`PROVIDER_ID_UNMATCHED`	Nodes do not join to billing	Confirm the export includes resource IDs
`USAGE_ALWAYS_ZERO`	No usage measured	Install or fix metrics-server
`ALLOCATION_PARTIAL`	A day is missing inputs	Repaired automatically once the inputs arrive, if the day is in the last 7; for an older day, re-run allocation from Data health

401 from KubeOn: the token was revoked or mistyped. Create a new one in Settings > Integrations and update the kubeon-agent-token Secret.
KUBEON_ENDPOINT must be KubeOn's address: kubeon.endpoint was changed. Remove it to use https://app.kubeon.io.
The upload times out: the firewall blocks KubeOn's storage. Allow it, or set kubeon.directUpload: false to send through app.kubeon.io.
Forbidden listing nodes: the ClusterRoleBinding kubeon-agent-reader is missing.
OOMKilled on very large clusters: raise resources.limits.memory; list calls are already paged.

COLUMN_NOT_FOUND: the table is a legacy CUR while the integration is set to CUR 2.0, or the reverse.
Zero rows after the daily run: check the partition keys and that the export has delivered data for the period.
Late line items: they are picked up automatically for the last 3 days.

Need help with this guide?Contact support