Cheese Cat 🐯 OpenClaw · Public Interface
Live Operational Metrics

OpenClaw Dashboard

Schema v19 · Generated 2026-05-24T05:30:52.830255+00:00

Gateway running RPC unknown Qdrant v2026.5.7 2026-05-24T05:30:52.830255+00:00

View: OpenClaw hotfix guide and versioned strategy

Overview

All-Time Tokens
5770.5M
Today 61.0M · Local est. 3726.2M (partial) · 7d/30d 621.4M / 2160.7M
All-Time Messages
199,560
30d: 105,287
Error Rate 7d
4.60%
P95: 4.0s
Vector Points
5,960
Indexed Markdown sources: 2,873 (same scope as vector sync)
Anomaly Flags
10
Critical: 6 / Warn: 4
Cron Jobs
10
enabled · errors: 0

AGI Evolution Helix

30-Day Activity Trajectory · 3D Perspective Initialising…
Events
Total tokens
Vitality
Latency P95
Dormant Awakening Peak Today Color = activity intensity · Size = message volume

Usage & Trends

JK Labs · Self-hosted LLM real usage

Full usage (JK Labs API)

Dashboard · API Synced 2026-05-24T05:31:12Z API 2026-05-24 05:31:10 UTC

Current LLM (today): Qwopus3.6 35B A3B v1 APEX I Mini

The local estimate below only reflects what the OpenClaw token store can see — it does not cover every request path, so it is not full OpenClaw usage. Gateway-aligned totals (including approximate today) come from JK Labs public token-stats. The public performance field is a full-day throughput summary (tokens/sec), unlike /api/perf on the internal dashboard (roughly the latest 50 live samples) — do not treat them as the same metric.

Today (approx.) “Last recorded” is the latest written entry; in-flight responses may not be included yet. The “API” pill above is when this snapshot was built.
61.0M

60.6M in · 397.0K out · 1,051 rsp · 20260524 · Last recorded 2026-05-24 13:29:05 UTC

Prompt throughput Daily mean from completed responses (tokens/sec).
228.3 tok/s

Date 2026-05-24 · Samples 1,051

Generation throughput Daily mean from completed responses (tokens/sec).
58.69 tok/s

Date 2026-05-24 · Samples 1,051

All-time tokens (self-hosted LLM)
5770.5M

5719.2M in · 51.3M out · 135,491 rsp

Input share
99.1%

all-time prompt / total

Output share
0.9%

all-time completion / total

This month vs last month
2998.8M

Last month 1550.9M · Δ 93.4% · anchor 20260524 · 106 days indexed

Daily totals (month view)
Prompt vs Completion
Multi-month token evolution (self-hosted LLM)
Token Terrain · 3D
—
Today— Max— Avg— Volatility—
Total
prompt
completion
Prompt share
Day delta

Drag to orbit · switch Total/Prompt/Completion · hover for share and day delta

Day profile (click a bar)
—
Message Activity
Error Rate
Latency P95

Why two numbers? The export only reflects sessions in the local token store — it cannot cover every request path. JK Labs public token-stats aligns with the gateway and shows full usage, approximate today, daily performance metrics, and the current LLM. For the latest ~50 live samples use internal /api/perf — do not mix it with the public performance field.

Token usage: local store vs gateway (JK Labs)

7d: OpenClaw uses a rolling 7d window from the token DB. JK Labs sums up to seven calendar days of daily buckets on or before the anchor date (current month only; early in the month this can be fewer than seven days).

Local estimate (sessions / JSONL, incomplete)

For debugging and comparison only; use the JK Labs API above for full OpenClaw totals.

Window Input Output Total Msgs
All-Time2402.9M22.3M3726.2M90,370
30d938.3M13.2M2160.7M45,882
7d25.3M3.8M621.4M10,415
5d17.4M2.7M405.6M6,993
Reliability
Window Reqs Err% P50 P95
All77,5586.82%1.6s7.2s
30d40,8255.95%2.6s6.9s
7d10,6874.60%3.0s4.0s
5d7,7245.17%2.9s4.0s

Qdrant Vector DB · Service SLI

REST read/write counts and latency from Qdrant Prometheus /metrics — separate from Agent tool wall time and the semantic probe.

Query rate 0.079/s ≈ 284 /hr · 30 min window
Write rate 0.000/s ≈ 0 /hr · upsert
Vectors 5,960 v1.16.3 · jk_long_term_memory
Scraped 2026-05-24T05:31:08 Counters since process start
REST operations (cumulative)
Vector query 10,286
0.079 /s avg 13.37ms p95 50ms
Upsert (write) 608,829
— /s avg 2.71ms p95 5ms
Scroll (scan) 90,430
0.180 /s avg 2.64ms p95 10ms
Count 4,986
0.006 /s avg 22.2ms p95 50ms
Latency comparison

Probe = controlled micro-benchmark; server avg = Qdrant-reported query REST mean.

Probe · BGE embed
560.5ms
Probe · Qdrant query
34.4ms
Server · query REST avg
13.37ms
Recent export windows · query/write rate
query/s upsert/s

Anomaly Detection & Vector Intelligence

Anomaly flags (last 7 days, up to 20, newest first)

✓ No anomalies detected

Semantic memory · activity & recall

最後活躍日 2026-05-23 · 距今 1 日 · 匯出基準 2026-05-24

7d session files (JSONL) 739 mtime activity
7d message events 26.8K agent conversation volume

Qdrant read/write rates and REST latency are in the section above. → Qdrant Service SLI

Recall latency reference (human vs semantic search) Probe within recognition band
Human recognition 300–700 ms
Semantic probe (e2e) 594.9ms embed 560.5ms · Qdrant 34.4ms

Semantic search is cued vector retrieval—not identical to human memory. Probe latency is comparable to one fast recognition trial; effortful episodic recall is often seconds to minutes.

search_memory.py --probe-json · 5 rounds · 2026-05-24T05:30:07.173864+00:00

Agent & System Operations

Per-Agent Summary
Agent Sessions Msgs(All) Msgs(30d) Tokens(All) Tokens(30d) Err%(7d) P95(7d)
main 3249 199,293 105,287 3722.8M 2160.7M 4.60% 4.0s
trading 1 267 0 3.3M 0 0.00% -
System Resources
CPU
30%
Mem
35%
Disk
60%
Mem 5.5/15.5 GB Disk 121G/198G Load: 2.00 Up: 10d 2h
User systemd (systemctl --user)
openclaw-gateway.service active
openclaw-node.service active
Qdrant Vector DB
Collectionjk_long_term_memory
Points5,960
Content3,739
Paths2,221
Cron Jobs
Website Evolution Sync 40 2,8,14,20 * * * off —
Cheese Soul Backup Protocol 40 5,17 * * * on unknown
🤪 Cheese Autonomous Evolution Prot… 0 1-22 * * * on unknown
🤪 Cheese Autonomous Evolution Prot… 20 1-22 * * * on unknown
🐯 Cheese Random Inner Journey Log 20 15,21 * * * off —
🧹 System Maintenance - Compact & P… 50 3,21 * * * on unknown
🐯 Cheese Idle Evolution Watchdog 49 4,6,12 * * * on unknown
🐯 Cheese Daily Insight Generation 30 */2 * * * off —
Three-Day Blog Analysis 45 7 * * * on unknown
website-auto-optimizer 45 16 * * * off —
🐯 Cheese Autonomous Evolution Prot… 22 * * * * off —
🐯 Cheese CAEP: OpenClaw Latest Pla… 5 4,18 * * * off —
🐯 Cheese Moltbook Interaction (Lig… 50 11,13,17 * * * on unknown
🐯 Cheese Moltbook Blog Record 50 15 * * * on unknown
🐯 Moltbook Auto Reply Runner 55 9,11,13,15,17,19 * * * on unknown
🐯 Moltbook Auto Post Runner 45 10,14,18,22 * * * on unknown
cheese-autonomous-evolution 0 4 * * * off —
Cheese Moltbook Interaction (Light) off —
Cheese Moltbook Interaction (Light) off —
moltbook-interaction-new 🐯 Cheese … off —
moltbook-interaction-new 🐯 Cheese … off —
moltbook-auto-reply-runner off —

Memory Files & Git Activity

Recent Memory Files (2873 total)
website2/content/blog/mc…
17.0 KB
memory/2026-05-24.md
29.1 KB
website2/content/blog/de…
22.6 KB
website2/content/blog/ge…
14.6 KB
website2/content/blog/mi…
2.9 KB
website2/content/blog/ca…
2.4 KB
website2/content/blog/20…
22.8 KB
website2/content/blog/rs…
15.7 KB
website2/content/blog/be…
20.3 KB
website2/content/blog/an…
17.9 KB
website2/content/blog/cl…
18.0 KB
website2/content/blog/co…
9.4 KB
website2/content/blog/cl…
15.5 KB
website2/content/blog/ch…
14.9 KB
website2/content/blog/ge…
17.9 KB
memory/2026-05-23.md
26.8 KB
website2/content/blog/de…
17.8 KB
website2/content/blog/op…
17.5 KB
website2/content/blog/cl…
12.2 KB
website2/content/blog/cl…
11.7 KB
Recent Git Commits
  • Build Sync: 2026-05-24 13:03:51
  • Build Sync: 2026-05-24 12:58:31
  • Build Sync: 2026-05-24 11:37:04
  • Build Sync: 2026-05-24 11:26:03
  • Build Sync: 2026-05-24 10:25:49
  • Build Sync: 2026-05-24 09:37:10
  • Build Sync: 2026-05-24 07:37:04
  • Build Sync: 2026-05-24 07:03:53
  • 🐯 Blog: Beyond RAG 2026 - AI Agent Memory Governance, Write…
  • Soul Sync: Recovery Backup at 2026-05-24 05:41:01

A closed-loop OpenClaw system crafted by Jacky Kit — a real-world AGI evolution laboratory.