New Relic APM Practitioner flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Apdex formula

Apdex equals Satisfied plus half of Tolerating, divided by Total transactions. Satisfied transactions are at or below T; Tolerating are between T and 4T. Default T for web transactions is 0.5 seconds.

Transaction naming precedence

Custom transaction names set in code or via API take priority over instrumentation-provided names such as framework routes or controller actions, which in turn beat the URI-based fallback.

APM 360 add-on

APM 360 is a separate paid SKU layered on APM that unlocks error profiles, service levels, and dependency/domain-level maps. Base APM alone does not include those capabilities.

Distributed trace headers

Distributed tracing propagates context using the W3C traceparent header, or New Relic's newrelic header when W3C is not enabled. Spans roll up into the transaction's trace view.

Key transaction

A key transaction is a single transaction flagged as its own entity for alerting and Apdex tracking, isolating it from the parent application's aggregate metrics.

Error rate calculation

Error rate equals error transactions divided by total transactions for the same time window. A transaction is counted once as an error even if it captures multiple exceptions.

Thread profiler purpose

The thread profiler samples JVM or .NET stacks to attribute CPU time to specific methods. It is not a heap analyzer; use JVM heap or GC metrics for memory pressure.

Core Web Vitals

The browser agent captures LCP (largest contentful paint), INP (interaction to next paint, which replaced FID), and CLS (cumulative layout shift) as part of real user monitoring.

Browser agent injection methods

The browser agent is added via a copy-paste JavaScript snippet, automatic APM injection for server-side apps, or a dedicated SPA framework agent for single-page applications.

Session replay

Session replay is an opt-in browser feature that records DOM mutations rather than video. It must be explicitly enabled and supports privacy masking of sensitive fields.

Synthetic monitor types

Synthetic monitors come in three types: ping (single HTTP request), simple browser (scripted steps), and scripted browser (Selenium-style). They run on a schedule from public or private locations.

RUM versus synthetic

Real user monitoring measures actual sessions and is subject to sampling. Synthetics are deterministic, run on a schedule, and fire regardless of whether real traffic exists.

Mobile agent telemetry

Mobile agents for Android, iOS, React Native, Flutter, and .NET MAUI report crashes, handled exceptions, network requests, and mobile vitals such as app launch time.

Page load timing phases

The browser agent reports page load timing as separate attributes: DNS lookup, TCP connect, TLS/SSL handshake, request and response, DOM processing, and page render.

New Relic AI assistant

New Relic AI is a natural-language interface over account telemetry. It answers questions and can generate NRQL, dashboards, and alert conditions, grounded in what data exists in the account.

AI Monitoring

AI Monitoring instruments LLM calls to capture model name, token counts, latency, cost, and prompt/response metadata for supported providers, enabling LLM observability.

Applied Intelligence

Applied Intelligence groups related alerts into incidents using correlation rules. It sits on top of alert policies rather than replacing them, reducing alert noise.

Anomaly detection

Anomaly detection uses baseline or seasonal models on metric timeseries, flagging deviations from the learned baseline rather than relying on fixed static thresholds.

AI-generated NRQL validation

AI-generated NRQL can be syntactically valid while referencing nonexistent attributes, so generated queries must be validated against the account schema before use.

AI Monitoring guardrails

AI Monitoring supports optional evals and guardrails for prompt and response safety when configured. Treat them as optional instrumentation layers, not defaults.

Infrastructure agent

The infrastructure agent is a host-level daemon on Linux or Windows that collects system metrics such as CPU, memory, disk, and network, and forwards logs and integration data.

On-host integrations

On-host integrations such as MySQL, NGINX, Apache, and Redis run as part of the infrastructure agent and report metrics under their own distinct event types.

Kubernetes integration components

Full Kubernetes coverage requires a DaemonSet for node-level metrics and a Deployment for cluster-level metrics and the kube-state-metrics scrape. Both are needed.

Cloud integrations model

Cloud integrations for AWS, Azure, and GCP are pull-based: New Relic polls the provider's APIs through a linked account or role, so no agent is installed on the resource.

Host not reporting condition

Host not reporting is a built-in condition type that fires when a host stops sending data for a configured threshold. It is separate from metric-based alert conditions.

Infrastructure agent config file

The infrastructure agent configuration lives in newrelic-infra.yml. Changes require a service restart, and logs go to /var/log/newrelic-infra/newrelic-infra.log on Linux by default.

Logs in context

Logs in context links log lines to APM transactions via trace and span IDs or entity GUIDs. It requires forwarding trace context through the log forwarder or agent.

Log forwarder

The log forwarder is Fluent Bit-based. It tails files, applies parsing rules, and ships lines to New Relic. The infrastructure agent can also forward logs directly.

Grok parsing rules

Grok patterns transform unstructured log lines into attributes. The parsed fields become queryable attributes in NRQL, enabling structured filtering and faceting.

Log retention default

Default log retention is 30 days, with extended retention available as a paid option. Log retention differs from metric and APM data retention periods.

NRQL clause order

NRQL clause order is SELECT, FROM, WHERE, FACET, TIMESERIES, then LIMIT. This order applies across all data types, including logs queried with FROM Log.

Alert policy flow

Alert policies contain conditions; conditions produce incidents; incidents are grouped by policy and routed to notification channels or workflows for delivery.