Instrumentation and Exporters Questions
Practice questions for Instrumentation and Exporters topic in Prometheus Certified Associate. 32 questions covering this domain.
A custom exporter collector updates global metric objects in place on every scrape. Why is that discouraged for non-trivial collectors?
A developer wants to expose how long it has been since the last successful sync. What is the recommended Prometheus approach?
A team proposes adding a `user_id` label to a high-volume request metric so every user's traffic can be queried directly. What is the Prometheus guida...
A service-level batch job runs nightly and then exits. Which metric is the key one to expose, and how should it usually be delivered?
An exporter can see application metrics plus machine CPU, memory, and filesystem metrics from the host. What should it normally do with the machine me...
A metric tracks the number of requests currently in progress and can move up or down during the day. Which Prometheus metric type fits best?
A batch job regularly runs for more than a few minutes and sometimes becomes slow when calling external systems. What additional monitoring approach i...
An exporter polls its target every 10 seconds on its own timer and caches the results for Prometheus to scrape later. What does the guidance say?
Which statement correctly describes Prometheus's default metrics collection model?
In Prometheus terminology, a "scrape target" is:
When should you prefer a Summary over a Histogram for measuring request latency?
What HTTP content type does a standard Prometheus metrics endpoint expose by default?
The Prometheus Node Exporter is primarily used to:
Which metric name suffix follows Prometheus naming conventions for a counter that tracks total HTTP requests?
The Prometheus Blackbox Exporter is used to:
Why is it harmful to use a `user_id` label with millions of distinct values in Prometheus?
A short-lived service-level batch job finishes before Prometheus can scrape it directly. What pattern do the docs recommend?
Why should instrumentation avoid labels such as `user_id` when they can grow very large?
An application currently exposes `http_responses_200_total` and `http_responses_500_total` as separate metric names. What redesign matches Prometheus ...
Which Prometheus metric type is appropriate for a cumulative total such as all processed requests since process start?
Sign in to see all 32 questions
Create a free account to browse all questions — completely free during our launch phase.