OpenClaw Observability Platform

English | 中文

OpenClaw Observability Platform, developed based on the KWeaver Core framework, uses OTel protocol and eBPF technology for full-link tracing and monitoring of AI Agents. It provides rapid fault diagnosis, security compliance management, and lean computing operations capabilities to ensure high-quality growth of AI-powered businesses.

Core Features & Business Value

24/7 Observability: Making OpenClaw Execution "White-Box"

• Core Capability: Build a comprehensive observation system providing lifecycle guarantees including pre-event (automated inspection), during-event (real-time monitoring & alerting), and post-event (precise fault diagnosis)
• Business Value (for IT Ops): Full-process transparency, eliminating black-box troubleshooting, ensuring 100% visibility and control of system status

Risk Perception: Enterprise-Grade "Brake System" for OpenClaw

• Core Capability: Establish robust security defenses covering real-time control (authorization management, compliance validation, storm blocking) and closed-loop auditing (audit traceability)
• Business Value (for CIO): Maintaining system security baseline, eliminating unauthorized calls and data security risks, achieving a perfect closed loop between business execution and security compliance

Productivity Assessment: Every Compute Investment Made Clear

• Core Capability: Based on multi-dimensional business accounting models, accurately decompose and track cost consumption across infrastructure computing, individual employees, and business departments
• Business Value (for CEO/CFO): Drive refined operations, reject "confusing compute accounts", and intuitively convert abstract LLM Tokens into clear business ROI

Architecture

┌─────────────────────────────────────────────────────────────────┐
│                    OpenClaw Observability Platform              │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│  ┌──────────────┐    ┌───────────────┐    ┌───────────────────┐ │
│  │   Frontend   │    │  Backend API  │    │  Apache Doris     │ │
│  │   (Vite+     │◄──►│  (Node.js)    │◄──►│  (OLAP Database)  │ │
│  │   React)     │    │  Port: 8787   │    │  Port: 9030       │ │
│  │  Port: 3000  │    └───────────────┘    └───────────────────┘ │
│  └──────────────┘                                               │
│                                                ▲                │
│                                                │                │
│  ┌─────────────────────────────────────────────┴───────────┐    │
│  │                  OTel  Data Pipeline                    │    │
│  │                                                         │    │
│  │  ┌─────────────┐   ┌──────────────┐   ┌───────────────┐ │    │
│  │  │   Sources   │──►│   Transform  │──►│    Sinks      │ │    │
│  │  │  (File/Exec)│   │(Remap/Reduce)│   │(HTTP to Doris)│ │    │
│  │  │             │   │              │   │               │ │    │
│  │  └─────────────┘   └──────────────┘   └───────────────┘ │    │
│  └─────────────────────────────────────────────────────────┘    │
│           ▲                                                     │
│           │                                                     │
│  ┌────────┴───────────────┐                                     │
│  │   OpenClaw Agent       │                                     │
│  │   Session Logs         │                                     │
│  │   (sessions.json /     │                                     │
│  │    *.jsonl)            │                                     │
│  └────────────────────────┘                                     │
└─────────────────────────────────────────────────────────────────┘

Core Components

Online Live Demo

Try it out now!

• URL: https://opsrobot-demo.aishu.cn:3000/

Quick Start

1. Environment Requirements

• Docker Desktop with Docker Compose plugin
• Node.js 18+

2. Clone the Project

git clone https://github.com/opsrobot-ai/opsrobot.git
cd opsrobot

3. Deploy Backend Services with Images

docker compose -f docker-compose.yml up -d

After services start, access: http://localhost:3000

4. Configure OpenClaw Data Collection

Note: Install and configure the Vector collector on each machine running OpenClaw. Vector Official Site | Vector Installation Guide

Vector Installation for macOS:

brew tap vectordotdev/brew && brew install vector

Vector Installation for Linux:

For CentOS:

bash -c "$(curl -L https://setup.vector.dev)"
sudo yum install vector

For Ubuntu:

bash -c "$(curl -L https://setup.vector.dev)"
sudo apt-get install vector

Modify vector.yaml collector configuration:

Vector Configuration Documentation

Point to the backend server IP address (if OpenClaw is on the same server, no modification needed):

sinks:
  session_to_doris: &sink_template
    uri: "http://127.0.0.1:8040/api/opsRobot/agent_sessions/_stream_load"

  session_logs_to_doris:
    uri: "http://127.0.0.1:8040/api/opsRobot/agent_sessions_logs/_stream_load"

  gateway_logs_to_doris:
    uri: "http://127.0.0.1:8040/api/opsRobot/gateway_logs/_stream_load"

  audit_logs_to_doris:
    uri: "http://127.0.0.1:8040/api/opsRobot/audit_logs/_stream_load"

  openclaw_config_to_doris:
    uri: "http://127.0.0.1:8040/api/opsRobot/openclaw_config/_stream_load"

  agent_models_to_doris:
    uri: "http://127.0.0.1:8040/api/opsRobot/agent_models/_stream_load"

Point to the actual OpenClaw log directory for log collection monitoring:

sources:
  sessions:
    command:
      - "sh"
      - "-c"
      - 'for f in ~/.openclaw/agents/*/sessions/sessions.json; do if [ -f "$$f" ]; then tr -d "\n" < "$$f"; echo ""; fi; done'

  session_logs:
    include:
      - "~/.openclaw/agents/*/sessions/*.jsonl"

  gateway_logs:
    include:
      - "~/.openclaw/logs/gateway.log"
      - "~/.openclaw/logs/gateway.err.log"

  audit_logs:
    include:
      - "~/.openclaw/logs/config-audit.jsonl"

  openclaw_config_file:
    command:
    - "sh"
    - "-c"
    - 'f="~/.openclaw/openclaw.json"; if [ -f "$$f" ]; then j=$$(tr -d "\n" < "$$f"); printf "{\"source_path\":\"%s\",\"openclaw_root\":%s}\n" "$$f" "$$j"; fi'

  agent_models_file:
    command:
    - "sh"
    - "-c"
    - 'for f in ~/.openclaw/agents/*/agent/models.json; do if [ -f "$$f" ]; then agent=$$(basename "$$(dirname "$$(dirname "$$f")")"); [ -z "$$agent" ] && continue; j=$$(tr -d "\n" < "$$f"); printf "{\"source_path\":\"%s\",\"agent_name\":\"%s\",\"models_root\":%s}\n" "$$f" "$$agent" "$$j"; fi; done'

  cron_jobs_config_file:
    type: exec
    command: 
      - "sh"
      - "-c"
      - 'for f in ~/.openclaw/cron/jobs.json; do if [ -f "$$f" ]; then tr -d "\n" < "$$f"; echo ""; fi; done'

  cron_runs_config_file:
    type: file
    include:
    - "~/.openclaw/cron/runs/*.jsonl"
    read_from: beginning
    fingerprint:
      strategy: device_and_inode

Start Vector Collector Service:

vector --config vector.yaml

5. Configure OpenClaw-Diagnostics-Otel Data Collection

• Official Documentation

In the openclaw.json file, add or modify the configuration as follows:

{
  "diagnostics": {
    "enabled": true,
    "otel": {
      "enabled": true,
      "endpoint": "http://192.168.72.87:4318",
      "traces": true,
      "metrics": true,
      "logs": true,
    },
    "cacheTrace": {
      "enabled": true,
      "includeMessages": true,
      "includePrompt": true,
      "includeSystem": true
    }
  },
  "plugins": {
    "entries": {
      "diagnostics-otel": {
        "enabled": true
      },
    },
    "allow": [
      "diagnostics-otel",
    ]
  }
}

After modifying the configuration, restart OpenClaw:

openclaw gateway restart

6. View All OpenClaw Observability Data:

• Interact with OpenClaw through its interface
• View collected data in the opsRobot product interface: http://localhost:3000

More Screenshots

Session Traceability Analysis:

Token Consumption Dashboard:

Digital Employee Module (Overview & Portrait):

Version Compatibility

This project closely follows the development of the OpenClaw community. It has been developed, validated, and tested based on the latest version of OpenClaw. For accurate collection and display of observability metrics, it is recommended to use in the following environment:

Contributing & Community

We welcome and encourage contributions in any form! Whether submitting bug reports, improving documentation, or submitting PRs for core code, all contributions are greatly appreciated to support the opsRobot open source community. Contributing Guide: Please read our CONTRIBUTING.md to learn how to get started. Community Exchange: WeChat group QR code

WeChat Community

Scan the QR code below to join the WeChat community:

License

Apache License 2.0