<@U04S8PLJWMT> hosting airbyte 2.0.1 on k8s, getti...
# replication-ask-ai
a
@kapa.ai hosting airbyte 2.0.1 on k8s, getting the error a few times a week. any issue on airbyte's side? Root Cause: The pod
replication-job-18438-attempt-0
failed to start due to a transient error when pulling a container image from the public Docker Hub registry. The issue resolved itself automatically after Kubernetes retried the image pull. Evidence: - The pod initially reported an
ImagePullBackOff
status, which triggered the alert. - Kubernetes events for the pod show the specific error:
Failed to pull image "airbyte/destination-bigquery:3.0.16": ... unexpected status from GET request to <https://registry-1.docker.io>
. - A subsequent check with
get_pod_details
confirmed the pod is now in a
Running
state, indicating the image pull eventually succeeded. Remediation: 1. No action is required. This was a transient, external issue, and the system's built-in retry mechanism worked as expected to resolve it. 2. The alert can be considered resolved.
k
There’s no indication in the provided knowledge that this is an Airbyte-side outage or systemic issue; the pattern matches transient Docker Hub / registry pull problems rather than a bug in Airbyte itself. In a similar case with
abctl install
, repeated
ImagePullBackOff
errors were traced to Docker Hub rate limiting and/or transient network issues when pulling Airbyte images, not to a problem with the images themselves. [ImagePullBackOff loop] Your description (intermittent failures, automatic recovery on retry, error coming from
registry-1.docker.io
) is consistent with that. Given the sources, the best-supported guidance is: • Treat these as transient registry/network issues unless they become persistent. • If they start happening frequently or pods stop recovering, investigate: ◦ Docker Hub rate limits and credentials (using authenticated pulls reduced
ImagePullBackOff
in the referenced case). [ImagePullBackOff loop] ◦ Network/proxy/firewall behavior between your cluster and Docker Hub. Based on the evidence you shared and the available docs, your “no action required, alert resolved” conclusion is reasonable. 👍 0 button 👎 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.