Prometheus Certified Associate study plan

Install Prometheus, node_exporter, and Alertmanager locally and verify each target reports UP on the targets page

Explain the pull model and contrast it with push-based collection in your own notes

Configure two scrape jobs with different scrape_interval values and confirm the difference in the targets UI

Describe the roles of TSDB, retrieval, and the HTTP API within the Prometheus server architecture

Write and validate PromQL selectors using label matchers, including negative and regex matchers

Apply range vectors with rates over counters and explain why rate needs at least two samples

Differentiate gauge, counter, and histogram behavior when queried

Break down a metric name into its namespace, subsystem, and unit conventions

Compute per-instance request rates and aggregate them with sum by and avg by

Use aggregation operators to collapse labels and predict the output label set before running the query

Diagnose and fix a query that returns no data due to label mismatch

Explain staleness, lookback delta, and how instant vectors resolve at a timestamp

Calculate error ratios using division with vector matching and on/ignoring clauses

Build histogram_quantile queries from bucket series and reason about bucket boundaries

Use offset, @ modifier, and subqueries to compare current and past windows

Identify which query patterns cause high cardinality or slow evaluation

Define the four golden signals and map each to a concrete metric in your stack

Instrument a small service with a Prometheus client library exposing a counter and a histogram

Compare exporters, pushgateway, and direct instrumentation and choose correctly for a batch job

Explain why labels must never carry unbounded values such as user IDs

Write recording rules that precompute an expensive query and verify the new series appear

Write alerting rules with for durations and labels, and confirm they fire in the Alertmanager UI

Configure an inhibition rule and a grouping route in Alertmanager

Explain how alert state transitions from pending to firing

Build a dashboard with panels for traffic, errors, latency, and saturation

Choose appropriate panel types and units for counters versus gauges versus histograms

Design alert thresholds that avoid flapping and explain the tradeoffs

Describe how distributed tracing and logs complement metrics in a debugging workflow

Review weak areas by re-deriving PromQL queries from memory without notes

Complete timed practice sets weighted toward PromQL and alerting

Explain federation, remote write, and service discovery at a conceptual level

Audit your stack for retention, cardinality, and configuration hygiene issues