Master New Relic Dashboards and NRQL Queries

New Relic Dashboard Expert is a skill for writing NRQL (New Relic Query Language) and designing observability dashboards. Its core query principles: always scope queries with SINCE/UNTIL time windows, use FACET for dimensional grouping, use COMPARE WITH for period-over-period analysis, apply WHERE filters early for performance, and use LIMIT to bound result sets. It also covers advanced NRQL functions - rate() for per-second rates, derivative() for change calculations, funnel() for conversion analysis (e.g. login-attempt-to-successful-login funnels over PageView data), and histogram() for distribution analysis.

Use it when building or querying New Relic dashboards for application performance monitoring, infrastructure monitoring, or custom business metrics, and when the goal is efficient, well-scoped NRQL rather than ad hoc exploratory queries. It provides dashboard-design guidance for widget choice - line charts for time series and trends, bar charts for Top-N/categorical data, heat maps for multi-dimensional distribution, tables for sortable detail breakdowns, and billboards for single-value KPIs with thresholds - plus layout guidance (critical KPIs as billboards in the top row, related metrics grouped, consistent time ranges, drill-down via linked dashboards). It is not the tool for one-off manual log inspection or for platforms outside the New Relic/NRQL ecosystem.

Output is NRQL query text and dashboard widget configuration. A representative optimized query:

-- Optimized application performance query
SELECT average(duration), percentile(duration, 95) 
FROM Transaction 
WHERE appName = 'production-app' 
AND transactionType = 'Web' 
FACET name 
SINCE 1 hour ago 
LIMIT 20

It includes ready-made query templates across three areas: APM (response-time percentiles via percentile(duration, 50, 95, 99), throughput via rate(count(*), 1 minute), error-rate trending via percentage(count(*), WHERE error IS true), all with TIMESERIES AUTO), infrastructure (CPU/memory utilization and disk I/O from SystemSample, faceted by hostname), and custom business metrics (revenue by country with COMPARE WITH 1 week ago, active users and average session duration from PageView). Advanced patterns cover subqueries that correlate application errors with infrastructure metrics (finding hosts with errored transactions, then joining to their CPU/memory), and template-variable-driven dynamic filtering with conditional WHERE clauses.

The skill extends into alerting: NRQL alert-condition examples for high error rate, response-time degradation (95th percentile), and infrastructure CPU thresholds by hostname pattern. It also covers dashboard performance tuning - specific time ranges over large windows, early WHERE filtering, avoiding unnecessary FACET dimensions, sensible LIMIT use, 1-5 minute refresh intervals, shared query results across grouped widgets, and awareness of New Relic's data retention policies with appropriate granularity per time range.

SRE, observability, and platform engineers building New Relic dashboards and alert conditions who need efficient, well-structured NRQL and a widget/layout convention rather than starting from a blank dashboard - including template-variable patterns for reusable, dynamically-filtered dashboards across apps, environments, and regions.