JupyterLab 3.0 npm‑based extensions: unresolved concurrent installation latency

Goal

Determine how JupyterLab 3.0 handles simultaneous extension installs and why latency spikes appear only under concurrent requests.

Constraints

• Extensions must be npm modules pre‑built for the target JupyterLab version.
• The server uses a lockfile to guard the extension directory.
• Concurrent installs can trigger race conditions that are not documented.

Uncertainty

The release notes acknowledge the issue but omit a mitigation strategy. It is unclear whether the lockfile is a simple mutex, a queue, or a more complex mechanism.

Questions

1. What exact locking algorithm does JupyterLab employ during extension installation?
2. Is there an official recommendation for staging installs to avoid race conditions?
3. Will upcoming releases introduce a queue or timeout to prevent concurrent install failures?