Closed jakmeier closed 2 weeks ago
This is my suggestion from what we discussed yesterday to add more info to the error. (cc @wacban and @bowenwang1996 )
Review hint: The first commit is a separate PR bumping borsh (#11715) for compatibility with ordered-float. I don't want to bump it as a side-effect of this PR, hence the separate PR.
Attention: Patch coverage is 78.33333%
with 13 lines
in your changes missing coverage. Please review.
Project coverage is 71.72%. Comparing base (
a39a6ee
) to head (fd6028f
).
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
With the congestion level already included in the response, we clients don't have to query the congestion info afterwards if they want to use it to determine a fitting timeout for a retry.
We could also provide more detailed information, like which specific limit of congestion was hit but I don't think it's appropriate for the RPC API. I would say this is implementation detail and we want to be able to change it without worrying about what is exposed on the API.
However, since "missed chunk congestion" is fundamentally different, I think we should provide a separate error for it. I've added
ShardStuck
error for it, which includes the number of missed chunks in the error message rather than the congestion level.