Closed mikeshultz closed 1 month ago
This might be a non-issue. Warning is here:
This is the graceful handling. My problem might be elsewhere. Will update when I learn more.
I'm closing this because it's already gracefully handling strings.
This might need to be better documented, the intent is to only allow metric data that can be plotted via the dashboarding system
The suggestion could be anything that can be turned into a label identifier such as a hex bytestring or address can be a key, so that is the recommended way you should capture things for plotting purposes
This may not actually be a bug, but I figure it's worth reporting anyway. Returning a dict from a task handler like
return {"foo": "bar"}
will fail, becausestr
is not an allowed return type (ScalarDatapoint
).Here:
https://github.com/ApeWorX/silverback/blob/2c6f2bc5a15e7b77c79add99d8493c8be760c4b3/silverback/types.py#L55-L60
Returning a
str
as a value causes an error on the entire task and results aren't handled. I wonder if there's some alternative here, like either stripping reference types, allowing these, or maybe failing a bit more gracefully.CC @fubuloubu