Skip to content

deviceid: stamp X-Device-UUID under /v2/harness-runtime as well as /v2/agents - #2071

Merged
SSharma-10 merged 1 commit into
mainfrom
imishra/harness-runtime-device-id
Oct 23, 2026
Merged

SSharma-10 merged 1 commit into
mainfrom
imishra/harness-runtime-device-id

Conversation

@mishraishika24-pixel

@mishraishika24-pixel mishraishika24-pixel commented Oct 23, 2026 •

Copy link
Copy Markdown
Collaborator

Summary

internal/deviceid stamps X-Device-UUID only on requests under the hosted agents base path. That path is moving from /v2/agents to /v2/harness-runtime, and without this change the header would silently stop being sent as soon as the client's base path changes, with no error anywhere.

This matches either prefix, keeping the path-segment boundary: /v2/agentsfoo and /v2/harness-runtimefoo are not caught, and neither is /v2/gen-ai/harness-runtime.

Safe to merge ahead of the godo bump

Unlike the godo and pydo changes, this is backward compatible: requests under /v2/agents still get the header, and requests under /v2/harness-runtime will too once the client moves. It does not change which URL doctl calls.

Follow-up (not in this PR)

Bumping godo to the release that uses /v2/harness-runtime is a separate change, after the API edge serves the new prefix.

Test plan

  • go test ./internal/deviceid/: passes.
  • Extended the in-scope table with /v2/harness-runtime and its sub-resources, and the out-of-scope table with /v2/harness-runtimefoo, /v2/harness-runtimefoo/ and /v2/gen-ai/harness-runtime.

🤖 Generated with Claude Code

…2/agents

The Harness Runtime API is moving from /v2/agents to /v2/harness-runtime. The
device-id transport stamps X-Device-UUID only on requests under the hosted
agents base path, so once the client's base path changes the header would
silently stop being sent.

Match either prefix, keeping the path-segment boundary so /v2/agentsfoo and
/v2/harness-runtimefoo are not caught. Safe to ship ahead of the godo bump:
today's requests still carry the header, and the new ones will too.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
@SSharma-10
SSharma-10 merged commit 10ba27c into main Oct 23, 2026
9 checks passed
@SSharma-10
SSharma-10 deleted the imishra/harness-runtime-device-id branch October 23, 2026 14:33
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants