Getting Started¶
TrueWatch provides a one-stop observability solution for cloud, cloud-native, applications, and business. Whether it is server status, application performance bottlenecks, or real user experience, you can collect data, perform correlation analysis, and gain real-time insights on a single platform. With TOBY AI TruePilot and TOBY AI Agents, you can get AI-assisted decision-making from anomaly detection to root cause identification, achieving a complete closed loop from observation to action.
TrueWatch offers two product editions, Commercial Plan and Deployment Plan:
Commercial Plan: designed for production environments, providing a complete feature stack and pay-as-you-go billing;
Deployment Plan: designed for private environments, requiring the TrueWatch system software to be installed on local infrastructure.
This document is intended for SaaS Commercial Plan users, guiding you through the complete closed loop from installing the collector to seeing your first monitoring data.
If you are using the Deployment Plan, refer to the private deployment documentation for installation and configuration.
Preparation: Register and Sign In¶
If you do not have a TrueWatch account yet, register a TrueWatch account first. After registration, select the site where your account resides to enter the workspace in the TrueWatch console.
Step 1: Install DataKit¶
DataKit is TrueWatch's official open-source data collection client. Deployed on your servers, it reports data such as host metrics, container status, and application logs to TrueWatch in real time. After installation, DataKit enables a set of basic collectors by default, so data collection starts without additional configuration.
TrueWatch supports multiple installation methods. You can choose based on your environment:
- Host installation: for operating systems such as Linux, Windows, and macOS
- Container installation: for Docker environments
- Kubernetes installation: for K8s clusters, with Helm deployment supported
- Other methods: including AWS ECS Fargate, AWS EKS, and offline installation
Go to the TrueWatch console, click Integrations > DataKit in the left navigation, select your operating system type, copy the installation command shown on the page, and run it on the target host. If the installation is successful, the terminal will display a success message.
For more installation details and advanced configuration, refer to DataKit Installation.
Step 2: Confirm Data Has Been Reported¶
After installation, DataKit enables a set of collectors commonly used on Linux hosts by default, so basic metrics are collected without manual configuration. Data is automatically reported to the workspace at each collector's default frequency, and you can usually see it in the console after a short wait.
To confirm that collection is working, go to Console > Infrastructure and check whether your server appears in the Host list. If the host name, IP address, and status display normally, DataKit has successfully connected and is working.
If the target machine does not appear in the host list, check the following:
- Check the DataKit service status: verify on the host that the DataKit process is running normally.
- Check network connectivity: confirm that the server can access the public internet, because DataKit needs to report data to TrueWatch.
- Check the workspace: confirm that the workspace currently selected in the console corresponds to the token used when installing DataKit.
If the issue persists, see DataKit No Data Troubleshooting for the complete troubleshooting workflow.
Step 3: Explore Built-in Visualization Views¶
After data is reported, the system comes with a series of built-in visualization view templates covering common scenarios such as hosts, containers, and databases. You do not need to build dashboards from scratch. Simply select and reuse these templates to quickly view the running status of your systems.
Go to Scenes > Built-in Views > System Views, select a view template under the Host category, and click it to view real-time charts of key metrics such as CPU usage, memory usage, disk I/O, and network traffic. You can create the template directly as a New Dashboard or New User View to apply it to your own workspace.
TrueWatch provides System Views templates covering infrastructure, middleware, application performance, cloud providers, and more, ready to use out of the box. System Views cannot be edited directly. To modify one, clone it to a User View first.
For more details, see Built-in Views and Quickly Create Dashboards.
Step 4: Configure Monitoring and Alerts¶
After confirming that data is being collected and visualized normally, we recommend configuring alert rules for your key metrics so that you are notified immediately when anomalies occur.
TrueWatch includes 700+ out-of-the-box monitor templates covering common scenarios such as host resources, databases, networks, and application performance. Go to Monitoring > Monitors > Create Monitor > Official Template Library, check "Host Detection Library" in the detection library on the left, select the detection rule you need (for example, host CPU average load too high), and create the monitor with one click.
After creating the monitor, configure the alert policy and notification targets. TrueWatch supports sending alerts through multiple channels, including email, DingTalk, Lark, WeCom, SMS, and Webhook, ensuring that alert information reaches the responsible personnel in a timely manner.
For details on monitor configuration and notification policies, see Get Started with Monitor.
Step 5: Explore More Data Types¶
After completing the basic host monitoring loop, you can gradually integrate more data types as needed to build full-stack observability:
| Data Type | Description | Recommended Scenarios |
|---|---|---|
| Logs | Collects application and system logs, with full-text search and pattern analysis | Troubleshooting and audit tracing |
| Application Performance Monitoring (APM) | Traces request links in distributed applications to identify performance bottlenecks | Microservices architecture and API performance optimization |
| Real User Monitoring (RUM) | Collects real user experience data | Web/mini-program performance optimization |
| Agent Monitoring | Tracks Agent requests and their correlation with the entire application trace, measuring the number of Tokens consumed by each generation task | AI applications and large model service governance |
| Security Monitoring | Integrates CSPM and SIEM capabilities to build a security monitoring system that unifies assets, configuration, and behavior | Security event management and compliance auditing |
| Cloud Resource Integration | Synchronizes resource data from cloud providers such as Alibaba Cloud, AWS, and Huawei Cloud | Unified monitoring for hybrid/multi-cloud environments |
Go to the Integrations module to view all data sources supported by TrueWatch and their corresponding integration guides. After all data is integrated, you can perform correlation analysis in the same workspace without switching between multiple tools.
Going Further: Query and Analyze Data¶
Once you are familiar with basic visualization and alerting, you can use DQL (Data Query Language), TrueWatch's self-developed query language, for more flexible data exploration. DQL is designed specifically for observability data such as time series, logs, and events, and is supported in many modules, including the Explorer, dashboards, Metrics Analysis, and the query tool.
For DQL syntax and usage examples, refer to DQL Query Language.
After completing the steps above, you now have what you need to use TrueWatch for basic system monitoring. To learn more about a specific feature, browse the corresponding documentation from the left navigation, or use the in-page help entry in the console for contextual guidance.
FAQ¶
Will installing DataKit affect server performance?
DataKit enables resource protection by default. When installed on the host machine, it limits CPU and memory usage through cgroups (Linux). Under low load, memory usage is about 100 MB and CPU consumption is low, so the impact on business systems is minimal. If the volume of collected data is large, you can adjust the [resource_limit] configuration in datakit.conf. For container or Kubernetes deployments, resource limits must be configured manually.
I no longer want to monitor a host. How do I remove it from the console?
After you stop or uninstall DataKit on the host, the host will no longer report data. If no data is reported for more than 24 hours, the host is automatically removed from the list. By default, the host list only displays hosts that have been online in the last 10 minutes. You can also click the "Only Show Online Hosts" button to view currently active hosts.
The names shown in the host list are IP addresses. Can I change them to custom names?
Yes. When installing DataKit, set the host tag through the environment variable DK_GLOBAL_HOST_TAGS (for example, host=my-server), or modify the global tag configuration in datakit.conf after installation. After the change, the console will display host names according to the tags you set.
Do I need to write code or complex query statements to configure a monitor?
No. TrueWatch provides an official template library covering common scenarios such as hosts, databases, and networks. You only need to select a template and adjust parameters such as thresholds and notification channels. For more flexible configuration, you can also customize detection rules through the visual interface without writing code.
Can I use TrueWatch to monitor my home computer, a virtual machine, or a Raspberry Pi?
Yes. DataKit supports multiple operating systems, including Linux, Windows, and macOS. As long as the device can access the public internet, you can install DataKit and connect it to TrueWatch. Virtual machines, cloud servers, physical machines, personal computers, and similar environments are all supported.
Where is my data stored? Is it secure?
For the SaaS Commercial Plan, data is uploaded and stored in the TrueWatch cloud data center. TrueWatch provides encrypted data transmission and storage mechanisms. If you have strict requirements for data sovereignty, you can choose the Deployment Plan to keep all data entirely in your local infrastructure.





