I’m developing a custom CloudQuery source and trying to sort out some of the error handling logic. My table resolver function is returning an error, and I see this in the go run main.go serve output. However, my cloudquery sync command claims to be completing successfully without errors (and the errors don’t appear in the cloudquery.log file).
I’m curious what the proper way is to surface these errors.
Hi @credible-calf,
If you’re running the plugin manually via
go run main.go serve
the errors will be shown in the console output.
In order to see them in the cloudquery.log, you can use
registry: local
and point
path:
to the absolute path of the compiled plugin binary.
It’s expected that a single table resolver error won’t stop the sync; otherwise, a timeout in a single endpoint can stop an entire sync, for example.
Thanks for clarifying that difference between the manual run and using registry: local. I’ll give that a try.
Will that cause these errors to appear somewhere in the cloudquery sync output and/or error count? I’m all for not having a single table resolver error stop the sync, but I do want to make sure we get some indication that an error occurred.
Yes, the errors will both appear in the cloudquery.log file (or console output if you sync with such options) as well as will be accounted towards the sync errors (reported at the end of the sync).
Perfect. Thanks a lot for clarifying. I’ll give that a try to confirm later.
I modified how I was running the source to use the registry: local approach. I can confirm that I now see the errors in the cloudquery.log file and the console output as well as in the status line.