Skip to main content
Every bot in MyClaw has a status that reflects its current state in the Kubernetes lifecycle. The status is shown on the dashboard bot list and on each bot’s detail page.

​
Status reference

StatusMeaning
pendingBot has been created in the database but its Kubernetes resources have not been scheduled yet.
deployingKubernetes resources (StatefulSet, Service, Ingress, Secret) have been created and the pod is starting up. OpenClaw is initializing and connecting to the gateway.
runningThe bot is fully started and connected to the MyClaw gateway. It is actively responding to messages.
stoppedThe bot has been manually stopped. Kubernetes resources remain but the pod is scaled to zero replicas. Data is preserved.
failedThe bot crashed or failed to start. Check the Logs tab for details. You can attempt a restart via the Start button.
deletingThe bot deletion has been initiated. Kubernetes resources are being removed. The PVC (data volume) is retained for 30 days.

​
Status transitions

pending → deploying → running
                    ↘ failed

running → stopped → running   (via Stop / Start)
running → deleting            (via Delete)
stopped → deleting            (via Delete)
failed  → deploying           (via Start)

​
What happens during each transition

pending → deploying MyClaw creates the Kubernetes Secret (encrypted config), StatefulSet, Service, and Ingress. The pod is scheduled on the cluster. deploying → running The OpenClaw container starts, completes its 5-step gateway handshake with MyClaw, and begins accepting messages. MyClaw’s StatusController watches the pod via the Kubernetes Watch API and updates the DB status on success. running → stopped MyClaw scales the StatefulSet replicas to 0. The pod is terminated gracefully. All config and conversation data on the PVC is preserved. any → failed The pod exited with a non-zero code or was killed by the liveness probe. Check Logs on the bot detail page to diagnose. any → deleting MyClaw removes the StatefulSet, Service, Ingress, and Secret immediately. The PVC is flagged for deletion after 30 days and cleaned up by a background job.

​
Troubleshooting failed bots

  1. Click the bot to open its detail page.
  2. Click the Logs tab to see the last container output.
  3. Common causes:
    • Invalid API key — check the key is correct and has not expired
    • Model name typo — verify the exact model identifier for your provider
    • Provider outage — try again in a few minutes
  4. Fix the config on the Config tab if needed, save, then click Start.