By symptom
| You see | Where it shows | Go to |
|---|---|---|
| A pod is crash-looping | Health → Pods counts it under crash-looping; the pod’s Restarts keep rising | Troubleshoot a Cluster → Read what the containers printed: read Previous instance on the pod’s Logs tab, then Logs |
| A pod stays Pending | Health → Pods counts it under pending; the pod’s phase reads Pending | Troubleshoot a Cluster → Find the object in Kubernetes: the pod’s Events say why it was not scheduled. When no node has room, add nodes in Cluster Nodes or turn on Cluster Autoscaling |
| A deploy failed | The cluster’s Technical details → Operations lists a failed operation | Find out why a deploy failed and retry it |
| The cluster is offline | The cluster shows Offline; Health says The cluster agent is offline. These are the last reported readings. | Ankra Agent → Troubleshooting, and what still works meanwhile |
| A cluster’s health check is not healthy | The cluster’s Health page shows a check or a score that is not healthy | Read a cluster’s overview and health → Go from a check that is not healthy |
| A hostname has no certificate, or does not resolve | The certificate stays pending; the Ingress host does not answer | Use Your Own Domain → What can go wrong |
| A push or pull request made no pipeline run | No Ankra pipeline check on the commit, or a check that says why nothing ran | Ankra CI Troubleshooting |
| The bill went up | The invoice or the usage report is higher than last period | Usage for which meter grew and on which cluster, Cloud Cost for each cluster’s spend, Cost Budgets and the Autopilot to cap it |
| A log search finds nothing older than the containers keep | Cluster logs → History offers to set up a log store | Ankra Hosted Logs, or connect a log source |
| No shell in the container | The Exec tab connects and the shell exits, or the image is distroless | Debug Pods |
A Secret’s values read [protected] | The Secret’s YAML tab | Kubernetes Resources → Secrets |
By message
Messages as the product shows them, with the page that explains each.| Message | Where | Explained on |
|---|---|---|
The cluster agent is offline. These are the last reported readings. | Cluster Health | Troubleshoot a Cluster |
The cluster is offline. These are previously recorded resources. | Kubernetes resource lists | Kubernetes Resources |
The cluster agent is offline, so no pods are reported. | Pods list | Troubleshoot a Cluster |
This object is managed. Change its source of truth. | Review resource actions | Kubernetes Resources → Change a resource |
Your role cannot perform this action. | Review resource actions | Kubernetes Resources → Change a resource |
Your role cannot read container logs on this cluster. | A pod’s Logs tab | Logs |
Live logs need the cluster agent. | Cluster logs → Live | Logs |
Ankra hosted logs are on, but no line has arrived from this cluster. | Cluster logs → History | Logs |
The log store could not confirm these lines are this cluster's. | Cluster logs → History | Log Sources → Sharing one log store between clusters |
This search could not be run. | Cluster logs → History | Logs |
This cluster's agent cannot send logs to Ankra yet. | An application’s Logs → History | Ankra Hosted Logs → What History tells you |
Connecting a log store of your own is not available | Settings → Integrations → Log store | Log Sources → What can go wrong |
You need kubernetes.exec permission on this cluster. | A pod’s Exec tab | Pod Terminal → Troubleshooting |
The terminal did not connect. Check the agent and try again. | A pod’s Exec tab | Pod Terminal → Troubleshooting |
AGENT_OUTDATED | ankra cluster debug create | Debug Pods → Requirements |
No trigger this pipeline declares matches this event, so there is nothing to run. | The Ankra pipeline check | Ankra CI Troubleshooting |
Awaiting authority approval | The Ankra pipeline check | Ankra CI Troubleshooting |
Still stuck
Press⌘J (Ctrl+J) on the page where you see the problem and ask Ask Ankra about it, or contact support.