How can I use the SeaLights Python Agent without modifying my repository?

Choose a deployment strategy when application source and Dockerfiles cannot change.

Problem

Use this approach when SeaLights references cannot be committed to the application source or Dockerfiles.

Solution

CI-managed injection keeps instrumentation changes in the delivery pipeline:

inject
  ↓
build
  ↓
test
  ↓
revert

The pipeline adds the Python startup hook and agent before testing. It removes temporary changes afterward. The application still needs SL_TOKEN, SL_BUILDSESSIONIDFILE, and SL_LABID.

This is a deployment strategy. The Python startup hook remains the runtime loading mechanism.

When to use it

Use CI-managed injection for:

• Regulated environments with strict repository controls.
• Shared product images owned by another team.
• Central platform teams running temporary coverage campaigns.

Choose the simplest option

Build-time installation is best when you own the image. Install the agent during the image build.

Kubernetes injection is best when you can change deployment manifests. Use an init container and shared volume.

CI-managed injection is best only when repository changes must be avoided.

Keep the SeaLights token in a Kubernetes Secret or your CI secret store. Do not commit it to source control.