Incident Management
Platform
AI
Dedicated compute
Dedicated compute keeps concurrent queries stable and complex analysis fast.
Learn moreDedicated compute
Dedicated compute keeps concurrent queries stable and complex analysis fast.
Learn moreWhen Copilot opens from an error log, failed trace, alert event, RUM error, or dashboard, it can use the current object and page state to offer a relevant starting prompt, reducing manual copying of fields and time ranges.
Ask about a log, trace, alert, RUM object, or dashboard in natural language. Answers depend on data and tools available under the user’s permissions; important conclusions should be checked against the query result and source records.
Copilot can write, explain, repair, or review a query and convert natural language into filters for the current Explorer. When drafting DQL, it checks workspace data sources, fields, tags, and syntax.
Describe the analysis goal and Copilot can produce a draft with charts and metadata. A user can revise the name, tags, visibility, charts, and variables, then confirm before saving it as a dashboard.
Continue from the current object and time range instead of moving disconnected snippets between tools.
Draft a readable DQL or PromQL query, then inspect its fields, filters, grouping, and window.
Queries, dashboards, and analysis content retain a human review step.
FAQ
Copilot is integrated into a Guance workspace and can use the current page context plus data and tools available under the user’s permissions. It does not bypass workspace access controls, and its output still requires review.
Current documentation covers logs, traces, alerts, RUM objects, and dashboards. Copilot can also answer product-documentation questions, assist with DQL or PromQL and Explorer filters, and draft dashboards or Pipeline parsing rules.
No. The documented flow creates a draft first. A user reviews and edits its metadata, charts, or variables and confirms before saving it as a dashboard.
No. Copilot can help form and investigate a hypothesis, but material conclusions should be verified against the generated query, filters, source telemetry, and relevant change records.
