Context
The published 57.1-second / 880 MB measurements belong to the removed WebContainer path and must not guide BrowserPod optimization. The adopted provider currently has no owner-authorized performance record.
Scope
Measure the exact BrowserPod 2.12.1 install and Gateway-ready path without changing behavior first. Treat metering, persistence, and transfer/storage cost as part of the result.
Acceptance criteria
- bind every measurement to the exact browser/OS, BrowserPod version, OpenClaw version, and npm integrity;
- capture cold install, warm install, persistent-storage reuse, Gateway-ready latency, installed bytes, cache/storage bytes, and failure/timeout behavior;
- separate provider boot/delivery time, npm install time, Gateway start time, and health/handshake time;
- run at least three samples per path and report median plus individual samples;
- verify that persistence invalidates on OpenClaw version or integrity change;
- redact credentials, tokens, portal identifiers where sensitive, and user payloads;
- add regression budgets only after evidence exists; do not carry forward WebContainer thresholds;
- run the normal report validation and affected browser evidence lane.
Context
The published 57.1-second / 880 MB measurements belong to the removed WebContainer path and must not guide BrowserPod optimization. The adopted provider currently has no owner-authorized performance record.
Scope
Measure the exact BrowserPod 2.12.1 install and Gateway-ready path without changing behavior first. Treat metering, persistence, and transfer/storage cost as part of the result.
Acceptance criteria