 |
VOOZH | about |
|
VOOZH | about |
You can launch a Bits investigation from several entry points:
You can invoke Bits on an individual monitor alert or warn event from several entry points:
Navigate to the monitor status page of a Bits Investigation-supported monitor and click Investigate with Bits Investigation in the top-right corner.
In the monitor event side panel of a Bits Investigation-supported monitor, click Investigate with Bits Investigation.
To use the Slack integration, connect your Slack workspace to Bits Investigation.
In Slack, reply to a monitor notification with @Datadog Investigate this alert.
When a Synthetic Browser or API test monitor triggers, you can launch a Bits investigation to identify the root cause. Bits analyzes Synthetic test results and history alongside traces, logs, and metrics. It surfaces a likely root cause and identifies whether the failure reflects a real regression or a misconfiguration.
The investigation opens in a new page, and you can also view it from the test details page after it runs.
Synthetic monitors support the same monitor-based entry points as other supported monitor types. See Monitor alerts for the available options, or toggle Auto-Investigate on a Synthetic monitor to start investigations automatically. For details, see Enable automatic investigations.
Click on New Investigation and describe the issue you want to troubleshoot. Include as much relevant context as possible:
The more specific your prompt, the more accurate and useful the investigation will be.
Good examples:
Bad example:
You can also trigger an investigation from Slack. Mention Datadog in a message: @Datadog Investigate high CPU in ai-gateway in prod over the last 30 minutes. If invoked within a Slack thread, Bits Investigation automatically uses the entire thread as investigation context.
In addition to manual investigations, you can configure Bits to run automatically when a monitor transitions to the alert state:
Bits is able to run investigations on the following monitor types:
APM Metrics type only; Trace Analytics is not supported)When Bits investigates an issue, it operates in a continuous loop of observation, reasoning, and action. It begins by forming hypotheses about the potential root cause, then uses its tools to query telemetry data to validate or invalidate those hypotheses. Each step builds on prior findings. As new evidence emerges, Bits updates its understanding, refines its reasoning, and chains together additional investigative steps—adapting and course-correcting until it converges on the most likely root cause.
At the end of an investigation, Bits either presents a clear, evidence-backed conclusion or marks the investigation as inconclusive when the available data is insufficient to support a defensible conclusion.
Bits uses the following data sources during investigations:
For best practices on maximizing the effectiveness of investigations, see Knowledge sources.
There are two display modes: Investigation Steps and Hypothesis Tree.
While an investigation is in progress, Bits captures every step it takes—including how it evaluates evidence and makes decisions—in the Investigation Steps view. This provides a real-time, detailed record of the agent’s reasoning process.
Once the investigation is complete, you can switch to the Hypothesis Tree view to explore a structured, tree-based visualization of the investigative path, making it easier to understand findings and conclusions at a glance.
The Reports tab enables you to track the number of investigations run over time by monitor, user, service, and team. You can also track the mean time to conclusion to assess the impact of Bits Investigation on your on-call efficiency.
Additional helpful documentation, links, and articles:
| |
