You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently we validate HTTP status codes outside of the pipeline (e.g., here. This means the retry policy has no way of retrying requests that failed by returning an HTTP status code that represents an error (e.g., 500s).
Some questions:
Should the retry policy inspect the incoming response and retry if the HTTP status code is 5xx?
Should the client immediately error (i.e., not retry) if the error is 4xx?
Should we validate the status code if it is 2xx or 3xx? For example, if the Cosmos create database endpoint returns 200 instead of 201, should we error? We currently will, but this seems very restrictive.
Currently we validate HTTP status codes outside of the pipeline (e.g., here. This means the retry policy has no way of retrying requests that failed by returning an HTTP status code that represents an error (e.g., 500s).
Some questions:
cc @heaths @MindFlavor
The text was updated successfully, but these errors were encountered: