Auto-Discovery

Chat automatically discovers components as it investigates production issues. You do not need to manually configure your status page upfront. Components appear as the assistant identifies distinct services, APIs, and features in your system.

How discovery works

When you ask Chat to investigate an issue, it analyzes logs, traces, and error patterns to understand what is happening. During this process, the assistant identifies specific services or features that are involved.

If the assistant determines that a service should be tracked on your status page, it can register it as a component:

1. Investigation begins: You ask about elevated errors or a specific problem
2. Analysis: The assistant queries logs and identifies the affected service
3. Discovery: The assistant recognizes this as a distinct, trackable component
4. Registration: The component is added to your project's status page

Once registered, the component appears on your status page and can have issues associated with it in future investigations.

Discovery triggers

Chat may discover and register components based on several signals:

Log patterns

When logs consistently reference a specific service name, the assistant recognizes it as a distinct component:

2024-01-15T10:23:45Z [checkout-service] ERROR: Payment validation failed
2024-01-15T10:23:46Z [checkout-service] ERROR: Database connection timeout

The repeated checkout-service reference indicates a trackable component.

Service mentions

During investigations, when the assistant identifies that a particular service is the source of problems, it may register that service. For example, if the assistant determines that errors are coming from your payment gateway, it will recognize payment-gateway as a component.

Error sources

When tracing errors back to their origin, the assistant identifies the responsible service. If that service is not yet registered, the assistant can add it as a new component.

No manual setup

The auto-discovery approach means:

• No upfront configuration: Your status page starts empty and grows organically
• Components reflect reality: Only services that actually appear in your logs are tracked
• Discovery is incremental: New services are added as the assistant encounters them
• Minimal maintenance: You do not need to keep a component list in sync with your architecture

This is particularly useful for microservices architectures where services change frequently. Your status page stays current without manual updates.

First seen vs last seen

Each component tracks two timestamps that help you understand its lifecycle:

First seen

The first seen timestamp records when the component was initially discovered and registered. This helps you understand:

• When the service first appeared in your system
• How long the service has been tracked
• The relative age of different components

Last seen

The last seen timestamp updates each time the assistant references the component during an investigation. This helps you identify:

• Active components: Recently referenced, likely still in use
• Stale components: Not seen in a long time, may be candidates for removal
• Lifecycle patterns: Services that are heavily used vs rarely encountered

Next steps