System Ops

Monitor instance health and manage runtime controls without touching the server directly.

URL: /admin/ops

Common tasks

• View instance metrics: CPU, memory, active chat sessions, and dev-server counts from the live runtime.
• Inspect logs: fetch recent API log output; secrets are automatically redacted before display.
• Rate-limit kill switch: temporarily disable or relax rate limits for the whole instance (useful during QA campaigns). Re-enable immediately after to restore normal protection.
• Manage allowlists / blocklists: add IP ranges or user agents to the egress allowlist or inbound blocklist.
• Copilot provider status: see which AI provider processes are running and their health state.

Rate-limit guidance

Warning

Only disable rate limits on a non-production environment, or for the shortest window needed. Restore limits as soon as the QA run ends. Never weaken other security controls as a workaround.

Log output is redacted for auth headers, JWTs, passwords, API keys, and database URLs. For unredacted output, SSH to the server directly.

See also: Users | Feature Flags | Signups