18% of the PCA exam. Sample questions below; the full library has 20 questions tagged to this domain.
Sample questions on Alerting & Dashboarding
Alerting & Dashboarding
Q1. How do multiple Alertmanager replicas coordinate to deliver notifications exactly once in a highly-available deployment?
Reveal answer and explanations
AA Raft-based leader election picks one instance to send all notifications on behalf of the cluster at any time during peer rotation
Incorrect. Alertmanager HA is not leader-based; Raft is used by other systems.
BThey share a Redis cluster that stores dispatch state, acquiring distributed locks per alert fingerprint before sending notifications
Incorrect. No external datastore like Redis is required or used.
CGossip mesh: peers share state, single delivery
Correct. Alertmanager peers connect via `--cluster.peer` and share state over a gossip protocol, so even though all peers receive alerts from Prometheus, only one sends the notification to each receiver.
DHA is not supported; you must run a single replica and fail over manually via a DNS change during any outage of the active node itself
Incorrect. HA via clustered gossip is explicitly supported and recommended for production.
Alerting & Dashboarding
Q2. Which of the following recording rule names most closely follows the recommended Prometheus naming convention `level:metric:operations`?
Reveal answer and explanations
A`http_requests_total_rate_5m`
Incorrect. This flat naming does not follow the `level:metric:operations` convention.
B`rate_5m_http_requests_total_by_job`
Incorrect. The order and punctuation do not match the colon-separated convention.
C`HTTPRequestsRate5m`
Incorrect. CamelCase is not part of the documented convention.
D`job:http_requests:rate5m`
Correct. `job:http_requests:rate5m` encodes the aggregation level (`job`), the metric name (`http_requests`), and the operation (`rate5m`), matching the recommended convention.
Alerting & Dashboarding
Q3. What is 'inhibition' in Alertmanager?
Reveal answer and explanations
AFiring alert suppresses notifications for matching alerts
Correct. Inhibition rules use source/target matchers so a higher-level alert (e.g. a cluster-wide outage) suppresses more specific alerts (e.g. individual probe failures) while it is firing.
BA rate-limiter that caps outgoing notifications per minute per receiver to protect downstream systems from bursts of pages
Incorrect. That describes throttling, not inhibition.
CA feature that restricts Prometheus rule evaluation to certain time windows configured globally in the rule manager schedule
Incorrect. Evaluation scheduling is a Prometheus concern.
DA PromQL function that masks out selected time series from query results based on label matcher expressions at query time
Incorrect. Inhibition is an Alertmanager configuration, not a PromQL function.
Alerting & Dashboarding
Q4. By default, when a firing alert's condition becomes false and the alert resolves, what does Alertmanager do?
Reveal answer and explanations
AIt does nothing; resolution is silent and only firing notifications are ever sent
BIt only sends resolved notifications when `for: 0s` is configured on the rule
Incorrect. Resolved sends are orthogonal to the `for` clause.
CResolved notifications are always suppressed if the alert was ever grouped with another alert
Incorrect. Grouping does not suppress resolved notifications on its own.
DIt sends a resolved notification to each receiver with `send_resolved` enabled (default for several integrations)
Correct. For receivers with `send_resolved: true` (the default for several integrations), Alertmanager sends a resolved notification when the alert clears, which is essential for on-call tooling state.
Alerting & Dashboarding
Q5. Which of the following notification integrations are available out-of-the-box in Alertmanager?
Reveal answer and explanations
AOnly email notifications with SMTP configuration and STARTTLS support for encrypted delivery to SMTP relays configured by the admin
Incorrect. Many receivers are supported natively.
BSlack, PagerDuty, email, webhook (and more)
Correct. Alertmanager ships with built-in receiver types for Slack, PagerDuty, email, OpsGenie, VictorOps, Pushover, Telegram, and a generic webhook that can bridge to anything else.
COnly the generic webhook — all other integrations require external plugins or sidecar processes to be deployed to the cluster hosts
Incorrect. Multiple receiver types are first-class, not plugin-only.
DNone — Alertmanager delegates all delivery to Prometheus, which runs the SMTP client and paging integrations itself internally via plugins
Incorrect. Prometheus forwards alerts to Alertmanager; it does not deliver notifications itself.
Alerting & Dashboarding accounts for 18% of the PCA exam. Expect questions that test recall of terminology and the ability to read short scenarios — not deep configuration. Use the sample questions above as difficulty calibration; if any feel hard, the rest of our 20-question domain bank will close those gaps.