[ci-fix] fix: add timeout to container runtime detection to prevent sonar system status from hanging

[github-actions]

Diagnosis

Category: Flaky / transient — test timeout caused by a missing subprocess timeout

The Integration and E2E Tests - linux job in Build run #27211240161 failed with a single test timeout:

(fail) system status > shows MCP configured when valid claude.json exists [15003.46ms]
  ^ this test timed out after 15000ms.

error: expect(received).toBe(expected)
Expected: 0
Received: 143   ← SIGTERM, process was killed by the test runner timeout

Root Cause

sonar system status calls checkMcpRunning() → detectContainerRuntime(), which probes Docker/Podman/Nerdctl by running <runtime> info via spawnProcess — with no timeout.

On CI runners where Docker is installed but the daemon is slow or unresponsive, docker info hangs indefinitely. The test runner kills the sonar process after 15 s (exit code 143 = SIGTERM), which is why result.exitCode is 143 instead of 0.

spawnProcessWithTimeout already exists in src/lib/process.ts for exactly this purpose.

Fix

Replace spawnProcess with spawnProcessWithTimeout(runtime, ['info'], {}, 3000, ...) in isRuntimeAvailable. A 3-second cap is more than enough for a healthy daemon to respond and prevents the hang on sluggish CI runners.

File changed: src/lib/tool-detector.ts

Triggering failure

• Workflow run: https://github.com/SonarSource/sonarqube-cli/actions/runs/27211240161
• Branch: master
• Commit: 9e881e84 by tomshafir-sonarsource — "CLI-640 Align and reorder sonar integrate install output"
• Failed job: Integration and E2E Tests - linux (job 80340803028)

Generated by Ci Failure Triage Agent · sonnet46 3M · ◷