I have two services on an App Engine project. Service A has a minimum 1 instance. Service B scales to 0. I would like when Service A receives a request, for Service B to receive a warm up request, or be initiated in some way. This is because users of the web application will primarily interact with Service A, and Service B just performs a couple heavy lifting tasks, which would be costly to have on all the time. However when the user is active, I would like to have Service B ready to go.
In other words, I need to Service B have an idle instance available only when Service A has been used recently. Is there a way to perform this?
The simplest way is to just use a dispatch.yaml:
https://cloud.google.com/appengine/docs/standard/python/reference/dispatch-yaml
Then, create a @before_request handler (it's a Flask thing) in Service A, and have it ping the Service B warmup url.
You could also create separate yaml files for each service, and have the url handler you want in serviceB.yaml (but not in serviceA.yaml).
https://cloud.google.com/appengine/docs/standard/python3/configuration-files
HINT: Make sure you deploy each of the newly-created yaml files using gcloud deploy