OpenClaw status performance note - 2026-05-21

Environment

• OpenClaw: 2026.5.19
• Node.js: 24.15.0
• macOS: 15.7.5 x64
• Gateway: loopback, running, connectivity probe ok
• Sessions: about 100 stored sessions across 2 active agent stores

Symptom

openclaw status appears to hang under short timeouts, but eventually exits 0.

Observed timings:

• openclaw status: about 42-49 seconds
• openclaw status --json: about 39-50 seconds
• openclaw gateway status: about 1-2 seconds
• openclaw channels status --probe: about 1 second

Findings

Gateway RPCs are fast and healthy. Log lines show status, channels.status, and doctor.memory.status completing normally.

The slow path is local status summary/session work. Direct module checks showed:

• collectStatusScanOverview(...): about 5 seconds
• scanStatus(...): about 46 seconds
• getStatusSummary(...): can take around 60 seconds cold

The expensive section is per-session model/runtime summary over stored sessions. Cleanup only found 2 missing transcript entries, so ordinary cleanup does not materially change the latency.

This appears distinct from #73531. That issue covers duplicate task reconciliation. In this environment the task registry is small, but status is still slow because session summary/model-runtime resolution dominates.

Expected

Default openclaw status should stay fast enough for routine health checks, ideally by summarizing only the recent N sessions or caching repeated model/runtime resolution.

Local workaround

Use:

scripts/openclaw-fast-health.sh

For full diagnostics:

scripts/openclaw-fast-health.sh --full