Using Azure features, how can we scale compute resources for an api/microservice prior to calling the API a) we need ability to specify infrastructure and performance features b) setup the API on that infrastructure c) then call the API while it is up/running for logging, status-ing d) shutdown on command that infrastructure
This sounds like it is scenario for Azure Batch.
Would this work? Create a linux image with the API setup, then use Azure Batch to create a pool specifying the image. There would not really be a task however I just need the image up/running so our application could submit to the image and API that was up and running.
Is this the best approach? (can you have a pool that starts without a real task and keeps up/running so we could execute an API that is exposed on it?)
Keep in mind, on the next API call we would repeat by creating a new pool with different compute requirements for the particular api call.