Deployment Stuck in "Provisioning"

A container that stays in the "Provisioning" state for more than a few minutes is usually blocked on one of three things: it can't pull the image, the region is temporarily constrained, or the health check never passes.

Symptoms

• The container remains in "Provisioning" well beyond the usual minute or two.
• No replicas become healthy, and no traffic is served.
• The container never reaches a running state, even though the configuration looks correct.

Likely causes

• The image cannot be pulled. The reference is wrong, the tag doesn't exist, or the image is private and no registry credentials are configured. See Image Cannot Be Pulled.
• Registry credentials aren't deployed to the container's regions. A registry must be present in every region the container runs in.
• The region is temporarily capacity-constrained. Provisioning can take longer during peak usage.
• The health check never passes. If the application never returns a 2xx on the health check path within the bootstrap time, the deployment can't be marked ready.

How to fix

1. Verify the image reference, including the registry host and the exact tag. Confirm the image exists by pulling it locally: docker pull <image>.
2. Check registry credentials if the image is private, and confirm the registry is deployed to the same regions as the container. See Registries.
3. Check the health check. Make sure the path exists in your application and returns a 2xx response. Raise the bootstrap time if the application is slow to initialise. See Health Checks.
4. Read the container logs if the application has started but isn't going healthy — a startup crash keeps a container from leaving provisioning. See Container Won't Start.
5. Give it a few minutes. Some deployments take longer during peak demand. If it remains stuck after that and the checks above are clean, contact support.

Related

• Image Cannot Be Pulled
• Container Won't Start
• Health Checks and Bootstrap Time
• Registries