Skip to content

fix: restore wait_for_http health check for interface-service before integration tests#8

Open
circleci-app[bot] wants to merge 3 commits into
develop-2from
chunk/fix-interface-service-health-check
Open

fix: restore wait_for_http health check for interface-service before integration tests#8
circleci-app[bot] wants to merge 3 commits into
develop-2from
chunk/fix-interface-service-health-check

Conversation

@circleci-app
Copy link
Copy Markdown

@circleci-app circleci-app Bot commented May 22, 2026

Summary

  • Restores the wait_for_http health check for interface-service (http://localhost:3567/health) before integration tests run
  • This check was accidentally removed in commit 489d7af ("trigger ci")

Root Cause

The user-service routes entity-management calls through interface-service (ENTITY_MANAGEMENT_SERVICE_BASE_URL=http://localhost:3567/entity-management). When integration tests run before interface-service is fully ready, user creation requests that involve entity-management hang indefinitely.

With no_output_timeout: 5m on the test step and mocha-junit-reporter writing output to an XML file (not stdout), CircleCI kills the process after 5 minutes of no console output — producing the empty test output seen in the CI failure.

Fix

Added wait_for_http for interface-service back in the correct position (after all other service health checks, before running integration tests):

- wait_for_http:
    label: interface-service
    url: http://localhost:3567/health

Test plan

  • Verify CI pipeline passes with this change
  • Confirm integration tests produce output and complete within the timeout

Original failing job

https://app.circleci.com/agents/circleci/3c4df9e5-2e07-4dba-99a6-6a869f03d07e/chat/e798ddf4-c407-4699-a9cd-0a3ce68523d9

circleci-app Bot and others added 3 commits May 22, 2026 16:19
… check in CI

The wait_for_http step for interface-service was polling http://localhost:3567/health,
but the interface-service has no /health endpoint of its own — it only proxies routes
to downstream services (e.g. /project/health, /user/health).

Hitting /health on port 3567 always returns a 404, which curl -sf treats as a failure,
causing the health check to exhaust all 60 retries and time out after 180 s.

Fix: change the URL to http://localhost:3567/project/health?basicCheck=true.
- /project/health is a registered route in the interface-service (via the elevate-project
  package loaded from elevate-routes.json) that proxies to the project-service on :5003.
- basicCheck=true skips the circular microservice dependency checks, consistent with the
  fix already applied to the direct project-service health check URL.
…vice-proxy-health-check

fix: use /project/health?basicCheck=true for interface-service health check in CI
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.

1 participant