PCA Practice Test 15 — 15 Questions

Question 1

A user executes an instant query for a metric without providing a range vector. What type of vector is returned by default?

Accepted Answer

Instant vector. An instant vector evaluates to a single data point per time series at the current evaluation timestamp.

Answer

String

Answer

Range vector

Answer

Scalar

Question 2

A Prometheus selector includes a negative regex match for environment names: 'environment!="staging"'. Which selector operator achieves this using regular expressions instead of exact matching?

Accepted Answer

environment!~"staging". The '!=' operator is for exact inequality, while '!~' is the negative regular expression match operator in PromQL.

Answer

environment!regex("staging")

Answer

environment!match("staging")

Answer

environment!="staging"

Question 3

A legacy batch job runs once per day and finishes in seconds. Why is a standard Prometheus pull model inadequate here?

Accepted Answer

The job is too fast for HTTP polling.. Prometheus pull intervals are typically set to seconds or minutes; a job that finishes quickly might never be 'up' when Prometheus scrapes it.

Answer

The data is too large for memory.

Answer

Pull model is insecure.

Answer

Prometheus cannot scrape batch jobs.

Question 4

Which TWO of the following are examples of metric data?

Accepted Answer

Current CPU usage percentage.. Metric data consists of numeric measurements captured over time, typically aggregated as counters, gauges, or histograms. Option B (Current CPU usage percentage) is correct because it is a numeric gauge value representing resource utilization at a point in time, exactly the kind of time-series measurement monitoring systems collect. Option D (Total number of completed HTTP requests) is correct because it is a numeric counter that monotonically increases as requests are served, a canonical metric used for rate and throughput calculations. Option A (a full database SQL query statement) is not metric data; it is a text/string artifact more suited to logs or traces. Option C (a line of text in an access log file) is log data, not a numeric measurement. Option E (the stack trace of an application error) is diagnostic trace/log data, not metric data.

Answer

A full database SQL query statement.

Answer

A line of text in an access log file.

Answer

The stack trace of an application error.

Question 5

What is the purpose of the 'instance' label in Prometheus?

Accepted Answer

To identify the target. It is an automatically added label that identifies the target (host:port) that the metric was scraped from.

Answer

To filter alerts

Answer

To identify the user

Answer

To identify the job

Question 6

Which TWO of the following are valid components of an Alertmanager configuration file?

Accepted Answer

receivers. Option A, receivers, is correct because the Alertmanager configuration file uses a top-level receivers: block to define notification integrations (for example email, Slack, PagerDuty, or webhook) that alerts are sent to. Option E, route, is correct because the top-level route: block defines the routing tree that matches incoming alerts and dispatches them to the appropriate receiver, including child routes and grouping/matching settings. Option B, storage_path, is not a valid Alertmanager configuration key; the data directory is set via the --storage.path command-line flag, not a config field. Option C, global_variables, is not a valid key; Alertmanager uses a global: block for settings such as SMTP and Slack defaults. Option D, scrape_configs, belongs to the Prometheus server configuration (defining scrape targets), not to Alertmanager.

Answer

storage_path

Answer

global_variables

Answer

scrape_configs

Question 7

When evaluating a rule with offset, such as http_requests_total offset 1h, what exact time does Prometheus query?

Accepted Answer

It evaluates the query at the evaluation timestamp minus 1 hour.. The offset modifier shifts the evaluation time backward by the specified duration relative to the evaluation timestamp.

Answer

It fetches data starting 1 hour in the future.

Answer

It calculates the moving average over the past 1 hour.

Answer

It delays rule evaluation by 1 hour to wait for late-arriving data.

Question 8

An SRE team is transitioning from traditional logs to metrics for performance monitoring. Which scenario best justifies using metrics over logs?

Accepted Answer

Tracking the request rate per second on a web server.. Metrics are numerical representations of data measured over time, making them efficient for time-series analysis and alerting, whereas logs are better for debugging specific events.

Answer

Auditing user access attempts for compliance.

Answer

Storing the payload of a failed transaction for later inspection.

Answer

Recording the stack trace of a Java application crash.

Question 9

Which monitoring philosophy is core to the Prometheus architecture?

Accepted Answer

Centralized pull-based scraping.. Prometheus favors the pull model, where the server controls the scrape rate and frequency, providing better control over the load on the monitoring system.

Answer

Agent-less log aggregation.

Answer

Push-based data ingestion.

Answer

Event-driven architecture.

Question 10

You notice your custom exporter is timing out during scraping. What is the most likely cause?

Accepted Answer

The scrape timeout is too short. If the exporter is overloaded or generating a massive amount of data, the scrape_timeout value in Prometheus might be too low.

Answer

The metric type is wrong

Answer

The port is blocked

Answer

The label names are incorrect

Question 11

When evaluating a histogram metric, what does the special pseudo-label le represent in bucket time series?

Accepted Answer

The upper bound of the cumulative histogram bucket.. The le (less than or equal to) label represents the upper bound of a histogram bucket.

Answer

The linear extrapolation factor for summary metrics.

Answer

The logging event identifier for tracing.

Answer

The exact latency error margin of the measurement.

Question 12

You have a recording rule that uses 'sum() by (instance)'. Why might this be inefficient for large clusters?

Accepted Answer

It can lead to high cardinality if 'instance' has many values. Using 'sum() by (instance)' in a recording rule aggregates metrics per instance, and if the 'instance' label has many unique values (e.g., in a large cluster with many nodes or pods), it creates a high number of time series. This high cardinality increases memory and storage usage in Prometheus, leading to inefficiency and potential performance degradation. The correct answer identifies high cardinality as the primary concern.

Answer

It blocks concurrent queries

Answer

It uses too much disk I/O

Answer

It slows down the network

Question 13

Which THREE of the following are required to expose a new metric in an application?

Accepted Answer

Register the metric. To expose a new metric in an application, you must first define the metric (E) by declaring it in code with a name, type (counter, gauge, histogram, or summary), and help text, since no metric can be collected without a declared instrument. Next, you must register the metric (C) with the application's metrics registry (e.g., Prometheus's default registry or a custom CollectorRegistry) so the instrumentation library knows to track and report it. Finally, you must expose an /metrics endpoint (D) over HTTP so the Prometheus server can scrape the registered metric values in the standard text exposition format. Enabling remote write (A) is a Prometheus server-side feature for forwarding scraped samples to a remote storage system and is not needed to expose a metric from the application itself. Configuring Alertmanager (B) is only for routing and sending alerts based on alerting rules, which is unrelated to the mechanics of exposing a metric.

Answer

Enable remote write

Answer

Configure Alertmanager

Question 14

Which THREE of the following are valid approaches to instrumenting an application?

Accepted Answer

Implementing a custom HTTP exporter. Option B is correct because implementing a custom HTTP exporter exposes application metrics on an HTTP endpoint (typically /metrics) in the Prometheus text exposition format, which Prometheus can scrape; this is a standard instrumentation pattern for applications whose metrics cannot be captured by existing client libraries. Option D is correct because Prometheus client libraries (e.g., Go, Java, Python, Ruby) are the primary recommended way to instrument application code, letting you define counters, gauges, histograms, and summaries that are registered and exposed automatically for scraping. Option E is correct because Pushgateway is the designated approach for short-lived jobs (batch/cron tasks) that cannot be scraped reliably; the job pushes its metrics to Pushgateway, which then exposes them for Prometheus to scrape. Option A is not valid because applications should never write directly into Prometheus's storage (TSDB); ingestion happens only via scraping or the remote-write protocol, and direct storage writes bypass the intended data model and lifecycle. Option C is not valid because running SQL queries against an application database is not an instrumentation mechanism; it does not produce Prometheus-format metrics or expose them for scraping, and it conflates data extraction with application instrumentation.

Answer

Directly writing to Prometheus storage

Answer

Using SQL queries to generate metrics

Question 15

What is the recommended approach for metrics that only need to be exposed periodically?

Accepted Answer

Expose consistently. The recommended approach is to expose metrics consistently, because Prometheus is a pull-based system that scrapes targets at fixed intervals regardless of whether the underlying value has changed. Consistently exposing the metric ensures that every scrape returns a current value, preserving continuity in the time series and enabling reliable rate(), delta(), and alerting calculations. Intermittent exposure creates gaps that break PromQL functions and trigger false staleness.

Answer

Only export when changing

Answer

Use Pushgateway

Answer

Delete the metric