Integrated Context
Agent0 Integrated Context automatically infers screen-specific details like filters, trace IDs, services, and time ranges when you open Search with Cmd+Shift+0, enabling hyper-relevant queries without manual input. Control context via eye icons, track origin pages for seamless navigation, and explore 13+ context types for traces, logs, metrics, and more. Reduces troubleshooting friction and accelerates insights directly from your Dash0 views.
Overview
Agent0 Integrated Context upgrades Search Cmd+Shift+0 to instantly capture your active screen's context—filters, span/trace IDs, services, time ranges, and more—eliminating the need to repeat details in queries. This makes troubleshooting conversational and precise, as Agent0 understands exactly what you're viewing, from traces to dashboards. Toggle contexts with eye icons and return to origin pages via "View {page}" buttons for fluid workflows.
Workflow Improvements
With Integrated Context, Agent0 automatically captures the key details of what you are looking at—such as the active service, filters, and time range—every time you open Agent0 Search. Instead of reconstructing the situation in a long prompt, you can now ask short, natural questions like “Why is latency high?” or “Why is this span failing?” and let Agent0 fill in the surrounding details. This makes analyzing your data feel closer to a conversation about the current view, rather than a repetitive description of it.
By removing manual steps like copying span IDs, recalling time windows, or retyping filters, this feature reduces friction and cognitive load during incident investigations. It helps you stay in the flow of debugging, quickly iterating on questions without losing the context of the screen you are working from.
Previous manual queries are streamlined:
| Previous Workflow | New Workflow |
|---|---|
| Specify service, time range, and filters manually in Agent0. | Open Agent0 Search Cmd+Shift+0 on any page; context auto-infers. Ask "Why is latency high?" |
| Copy span/trace IDs and details before querying. | Select span on Tracing page, hit Cmd+Shift+0, ask "Why is this failing?"—IDs included automatically. |
This cuts steps, boosts accuracy, and supports focused time ranges from Outlier maps or charts.
Controlling Inferred Context
Sometimes you need to ask broader questions that ignore specific active filters or screen context. To exclude an inferred context from your query, click the Eye icon next to the value in the Agent0 Search panel.
The tag grays out and the icon changes to a crossed-out eye, confirming Agent0 temporarily ignores that context for your current search. This gives precise control, letting you mix inferred details with wider scopes without resetting your view.
Experience smarter Agent0 context today—hit Cmd+Shift+0 and ask your first contextual question!
Sometimes, however, you might want to ask a broader question that *ignores* some of the active filters or context. If you wish to exclude a piece of inferred context from your query, press the **Eye icon** next to the inferred value within the Agent0 Search panel. **The tag will grey out and the icon will change to a crossed-out eye** to confirm that Agent0 is temporarily ignoring that specific context for your current search.
Origin URL Navigation
When you start a new Agent0 conversation from a page, that origin is stored and displayed at the end of the thread as a “View {page}” button. This allows you to jump straight back to the exact screen where you first opened Agent0, without having to retrace your steps or rebuild filters and views manually.
This feature closes the loop between investigation and action by letting you explore Agent0’s insights in depth and then return seamlessly to your original context. It reduces disorientation, keeps your mental model aligned with what you were analyzing, and helps Agent0 feel like an integrated part of your debugging workflow rather than a separate tool.
Full Context Types
Agent0 recognizes these for precise scoping:
| Context Type | What it Represents | Usage Example |
|---|---|---|
filters | Active key/value filters | Constrains to current view like service.name: "checkout-service". Ask "Why high latency with these filters?" to analyze filtered telemetry without retyping conditions. |
time_range | Current from/to timestamps | Limits analysis to your visible period. Query "Show error trends" on a zoomed chart for period-specific insights. |
service | Service name/namespace | Scopes to "Show errors for this service." Use on service overview: "Compare latency to baselines" for component-focused diagnostics. |
span | Selected span/trace ID | Deep dives like "Analyze this operation." Select span and ask "What caused this delay?" for root cause in distributed transactions. |
log | Highlighted log ID | Queries log content and links. Highlight log and type "Related traces?" to find associated telemetry. |
metric | Selected metric name/type | Chart questions like "Error rate trends?" Click metric chart: "Forecast next hour?" for predictive analysis. |
promql | Query string | Explains or optimizes PromQL. View visualization: "Simplify this query" for cleaner expressions. |
dashboard | Dashboard/panel ID | "Improve this panel." On dashboard: "Add error rate alert" to automate configurations. |
failed_check | Alert/issue ID | "Why did this check fail?" From alerts page: "Suggest fixes" for proactive remediation. |
document | External URL/title | Cross-reference docs. Viewing docs: "How does this apply to my trace?" for integrated knowledge. |
website | Website monitoring | "Latency for this site?" On website view: "Core web vitals breakdown" for performance details. |
web_events | User actions | "Errors for this event." Select event: "User impact?" to quantify session errors. |
session | User session | "Full flow analysis." Replay session: "Bottlenecks in checkout?" for journey debugging. |
Last updated: December 11, 2025
