Skip to content
Runway Docs

Incident Response

This page outlines processes that help with incident response for Runway services.

Restart Service

Cloud Run does not allow us to restart individual or all instances of a service. If you need to restart all instances, e.g. to purge data cached in memory, you need to deploy a new “dummy” release. See Emergency Deployment.

Emergency Deployment

Runway supports multiple Deployment Strategies. We recommend to use the normal deployment strategy when possible, since it is the best tested and most familiar process.

That said, there are cases when the default deployemnt process is broken or too slow. In such situations the expedited deployment strategy can help to deploy a new version as quickly as possible.

To enable the expedited deployment strategy, add the following to the .gitlab-ci.yml file of your service repo:

variables:
  RUNWAY_DEPLOYMENT_STRATEGY: expedited

Adding the variable via .gitlab-ci.yml has several advantages:

  • The emergency deployment is triggered by the push.
  • The setting is more easily discoverable.
  • There is a paper trail who added the variable and when.