Know when a Dispatched Action Finishes

[rlove]

Select Topic Area

Product Feedback

Body

As Deploys grow in complexity in our Enterprise, we have found that we orchestrate deployments across multiple repositories. We also need to watch the status of all of our executed deployments from multiple repositories so we can have a single dashboard of all of the Deployments that are executing. This needs to be segregated by teams and the repositories that they own.

The following feature request would allow us an elegant way to take advantage of this.

Create a new version of "CreateDispatch" API that returns the run_id. If this is not possible due to a distributed architecture then have it return a unique token/guid, if successfully started. Then we would need an additional FindRun(owner,repo,token/guid) api call. Yes, many people could implement this wrong and run out of API call limits quickly. So, another option would be to establish a Web Hook that will trigger with a payload of the token/guid.

But we know that our deploys take ~4 minutes to spin up and complete. So, the call to "FindRun" would typically not need to first occur until the 4-minute window has passed. Many other strategies could be used to find ways to hitting the API limit. Although the API limit of enterprise customers we would have really try hard to blow through our limits.

[rlove]

This is now implemented