fix: add timeout to health check exec calls to prevent blocking

[claude-claude]

CI Fix

Fixes CI #21696494896

Problem

The test was timing out on ARM64 because the VM never became healthy within the 120-second timeout. The health monitor was getting blocked when trying to run fcvm exec to check container status via podman inspect.

Root cause: Commit cb639e9 changed the health check to not assume healthy when exec fails, which is correct behavior. However, the fcvm exec command has built-in retry logic with exponential backoff that can take 50+ seconds when the exec server isn't ready yet (common during VM startup). This caused health checks to hang for extended periods, preventing VMs from being marked as healthy.

Solution

Added a 2-second timeout to exec calls in both check_container_running and check_podman_healthcheck. This allows the health monitor to fail fast and retry on the next iteration (every 100ms during startup), rather than blocking for the full exec retry duration.

With this fix:

• Health checks complete in <2s even if exec isn't ready
• The health monitor can poll frequently without blocking
• VMs become healthy quickly once the exec server is ready
• No changes to the correctness of the health check logic

---

Generated by Claude | Fix Run

[ejc3]

Already fixed in http-proxy branch (commit cb639e9 and 4963268 include these changes).