feat(otlp-grpc): add retry logic and comprehensive error handling to … - #2076
feat(otlp-grpc): add retry logic and comprehensive error handling to …#2076arjun-rajappa wants to merge 6 commits into
Conversation
…trace exporter - Implement exponential backoff retry mechanism for transient gRPC errors - Add retry support for Unavailable, DeadlineExceeded, Cancelled, ResourceExhausted, Aborted, Internal, and DataLoss errors - Improve certificate handling by reading file contents instead of passing file paths - Add timeout support with deadline tracking across retry attempts - Implement metrics reporting for export failures - Add comprehensive test coverage for error scenarios, retry logic, and edge cases Signed-off-by: Arjun Rajappa <arjun.rajappa@ibm.com>
|
👋 This pull request has been marked as stale because it has been open with no activity. You can: comment on the issue or remove the stale label to hold stale off for a while, add the |
Signed-off-by: Arjun Rajappa <arjun.rajappa@ibm.com>
…ILURE, and RETRY_COUNT Signed-off-by: Arjun Rajappa <arjun.rajappa@ibm.com>
kaylareopelle
left a comment
There was a problem hiding this comment.
Thank you for adding all the defensive error handling to the gRPC exporter! I have a few questions about the tests.
- Replace all wont_be_nil assertions in #initialize tests with specific equality checks (@timeout, @shutdown, @host) so each test validates the exact value that was configured - Rewrite 'encodes dropped attributes, events, and links counts' to intercept the proto request and assert dropped_attributes_count, dropped_events_count, and dropped_links_count directly on the span proto, rather than only checking the export return value - Rewrite 'translates all the things' to mirror the pattern in otlp-common: create spans via a real TracerProvider using with_ids to pin IDs, then build a complete ExportTraceServiceRequest proto and assert equality in a single must_equal rather than field-by-field - Consolidate TRACING_INTEGRATION_TEST guard from individual skip calls into a single before block on the 'integration with tracer provider' and 'performance' describe groups Signed-off-by: Arjun Rajappa <arjun.rajappa@ibm.com>
|
@arjun-rajappa - Can you take a look at the merge conflicts in this PR? |
|
👋 This pull request has been marked as stale because it has been open with no activity. You can: comment on the issue or remove the stale label to hold stale off for a while, add the |
|
Hi @arjun-rajappa, thanks for fixing the merge conflicts. We still have a few failing checks. Can you take a look at those? |
Closes #1667
Changes