paulononato/backstage
1
0
Fork 0
Code Issues Pull Requests Actions 47 Packages Projects Releases Wiki Activity

fix(integration): correct GitLab fetch mode and retry on network errors

Browse Source 
Removes a misplaced `mode: 'same-origin'` option that would have rejected
cross-origin requests when the integration is used from a browser, and
extends the retry wrapper so transient network errors are retried using
the configured `maxRetries`. Caller-initiated aborts still propagate
immediately.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Signed-off-by: Fredrik Adelöw <freben@spotify.com>
This commit is contained in:
Fredrik Adelöw
2026-05-04 10:43:49 +02:00
parent e5f31ff3a2
commit 6b112d3fe8
3 changed files with 93 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.changeset/gitlab-integration-fetch-fixes.md
+8
View File
@@ -0,0 +1,8 @@
---
'@backstage/integration': patch
---
Fixed two issues in the GitLab integration's fetch behavior:
- The internal fetch wrapper was passing `mode: 'same-origin'` on every request. This had no practical effect server-side, but would have caused cross-origin requests to be rejected when the integration is used from a browser. Requests now use the default fetch mode and work correctly in both browser and Node environments.
- When retries are configured, transient network errors (such as dropped connections or DNS hiccups) are now retried using the same `maxRetries` and exponential delay as retryable HTTP status codes. Previously, a thrown fetch error would propagate immediately on the first failure regardless of the retry configuration. Caller-initiated aborts continue to surface immediately without being retried.