Only Restart Affected Containers

[abhibongale]

Optimize rebuild to only restart containers whose images have actually changed. If only Ironic code changed, don't restart infrastructure services. This reduces rebuild time and avoids unnecessary service interruptions.

Goals

• Detect which images changed
• Restart only affected containers
• Skip unnecessary restarts
• Improve rebuild reliability

Implementation

Add image change detection to stackbox/core/builder.py:

def get_image_id(tag: str) -> Optional[str]:
    """Get image ID for a tag."""
    result = subprocess.run(
        ["docker", "image", "inspect", tag, "--format", "{{.Id}}"],
        capture_output=True,
        text=True
    )
    return result.stdout.strip() if result.returncode == 0 else None


def image_changed(tag: str, old_id: str) -> bool:
    """Check if image changed."""
    new_id = get_image_id(tag)
    return new_id != old_id

Update rebuild command to track image IDs and only restart if changed.

Testing

# Make code change
stackbox rebuild ironic
# Should restart ironic containers

# No changes, rebuild
stackbox rebuild ironic
# Should detect no changes, skip restart

Acceptance Criteria

• Detects image changes
• Skips restart if no change
• Restarts only affected services
• Shows what was restarted

Resources

• [Docker image inspect](https://docs.docker.com/engine/reference/commandline/image_inspect/)