fix(bigquery/storage/managedwriter): retry improvements #9642
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR makes two changes to retry behaviors in managedwriter.
In the first, this PR expands the set of conditions that trigger reconnect when sending the initial request to the backend.
In the second, this PR adds some additional handling for context cancellations when reading responses back from the service. In cases like reconnection, we establish a new Connection, each of which has it's own associated context. When draining remaining writes from a connection that's being shut down, we now pass the write into a retryer with a status-based error rather than raw context.Canceled, so we can recover more cleanly if the user is leveraging write retries.
Related internal issue:
b/326242484