Skip to content

originalRequestId propagation in executeRequest #2163

Description

@coderabbitai

Description

originalRequestId not initialized on first retry

The originalRequestId parameter is always nil on the first call from executeAblyRequest:...fallbacks:retries:originalRequestId:. This causes:

  • When _options.addRequestIds is enabled and fallbacks are used, the first fallback attempt has originalRequestId == nil, potentially appending a request_id query item with a nil value on top of the original request's ID.
  • Errors produced on fallback attempts will have a requestId of nil, even though the original attempt had a concrete requestId, weakening correlation and debuggability across retry attempts.

Expected behavior: All fallback attempts should share the same request_id value from the original request, and propagate a non-nil requestId into ARTErrorInfo instances.

Context

Additional Information

This issue exists in the current codebase and is not introduced by PR #2158.

┆Issue is synchronized with this Jira Task by Unito

Metadata

Metadata

Assignees

Labels

bugSomething isn't working. It's clear that this does need to be fixed.

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions