potassium Requests now have these fields:
.id - the request ID from Banana, sent as "X-Banana-Request-Id" in headers
.headers - the headers from the original request
Why?
.id is needed so users can include the request.id in their logs for filtering in Banana
.headers is needed because we're an important http framework 💪
How did you test it works without regressions?
Tests pass, added new tests.
If this is a new feature what may a critical error look like?
Headers, depending on the http client sending them, may have their cases changed, so the assumption is that users test their code before using this
Things to consider to not repeat mistakes we've learned from many times
[ ] If critical errors fire do we ping team in some obvious way (e.g., slack)?
[x] Are there debug logs + a way to see these logs if we need to debug?
[x] Is this documented enough a dev could work on this code without getting stuck or having to ping you?
What is this?
potassium Requests now have these fields: .id - the request ID from Banana, sent as "X-Banana-Request-Id" in headers .headers - the headers from the original request
Why?
.id is needed so users can include the request.id in their logs for filtering in Banana .headers is needed because we're an important http framework 💪
How did you test it works without regressions?
Tests pass, added new tests.
If this is a new feature what may a critical error look like?
Headers, depending on the http client sending them, may have their cases changed, so the assumption is that users test their code before using this
Things to consider to not repeat mistakes we've learned from many times