Get the NRQL answer without rebuilding the query

Ask about a spike and get the numbers back

Listing APM applications first is how a query gets scoped by the real application name instead of the one you half-remembered. Then the NRQL runs and comes back with results plus metadata for event types, facets and the window that was actually queried.

The dashboard tools solve the same problem from the other end. Fetching a dashboard returns its pages and widgets with the NRQL exposed in each widget’s configuration, so a companion reads how a query somebody already trusted was written and adapts it rather than composing one from scratch.

Why is one query tool most of the product?

Because NRQL is the interface to New Relic rather than a convenience layer over it.

Metrics, logs, traces, custom events and APM data all arrive as query results, so a single tool running arbitrary NRQL reaches further than a dozen narrow endpoints, and the other five exist to make that one usable.

Creating alert conditions, editing dashboards, acknowledging or closing incidents, managing workloads and posting deployment markers have no tools here. Those are handful-of-times-a-month decisions, and they stay with you.

What does a companion do every morning here?

Runs the loop nobody has time for.

The incident list, the NRQL that explains one, the dashboard somebody already built for that service, and a written answer at the end. Eight tabs turn into three paragraphs before standup.

Credentials are a numeric account id, a User API key beginning NRAK, and optionally a region of US or EU. The key inherits the permissions of the person it belongs to, so a read-limited user is a sufficient answer here. The one cost to watch is an unbounded NRQL query over a large account, which is expensive on New Relic’s side and spends credits on ours.

Write the incident note that names the cause

The spike lines up with a deploy in another dashboard, a support thread that started an hour earlier, and a vendor status page. What people want is a note naming all four, and a companion writes it by reading across the accounts you are signed into with the NRQL as one input. That is research shaped like an incident review.

Turning a month of those notes into a pattern is data extraction: which services page most, which alerts fire three times for one cause, which incidents traced back to a deploy. New Relic ships its own in-platform assistant, as of this writing, sitting on the telemetry inside a single account.

Frequently asked questions