Skip to main content

Code location deployment fails due to container configuration

Problem description​

Users may encounter deployment failures when attempting to deploy code locations with container configurations in Dagster+. This issue can occur even when the same configuration works in other environments, typically manifesting as an error related to container limits.

Symptoms​

  • Code location deployment fails in one environment but works in another.
  • Error messages related to container configuration or resource limits.
  • Deployment failures persist even after adjusting resource limits in the code location configuration.

Root cause​

The issue occurs when the Dagster+ agent configuration is missing the required containerConfig key in the YAML configuration, even if there are no specific values set for it. This prevents proper handling of container configurations specified in code locations.

Solution​

Add the containerConfig key to your Dagster+ agent configuration YAML, even if empty:

serverK8sConfig:
  containerConfig: # raw config for the pod's main container
  podSpecConfig: # raw config for the spec of the launched pod
    nodeSelector:

Step-by-step resolution​

  1. Locate your Dagster+ agent configuration file.
  2. Add the containerConfig key under serverK8sConfig as shown in the example above.
  3. Redeploy your Dagster+ agent with the updated configuration.
  4. Attempt to deploy your code location again with the desired container configuration.

Alternative solutions​

If issues persist, verify:

  • Kubernetes resource quotas in your environment
  • Node resources availability
  • Matching Dagster versions across environments

Prevention​

Always include the containerConfig key in your Dagster+ agent configuration when working with containerized deployments, even if you don't need to specify any values for it initially.