PCA Federation & Remote Storage Flashcards

In cross-service federation, which metrics are typically federated up to a global Prometheus server?

Answer: Pre-aggregated or recording-rule-generated metrics relevant at the global level

Cross-service federation is used to pull a curated subset of high-level metrics (often generated by recording rules) rather than all raw data, keeping the global Prometheus lean.

What happens to 'job' and 'instance' labels when metrics are federated from one Prometheus server to another?

Answer: They are preserved as-is from the originating Prometheus server

When Prometheus federates metrics, it preserves the original labels including job and instance, because honor_labels defaults to true in typical federation scrape configs.

Which open-source project extends Prometheus with global query capability and long-term object storage integration?

Answer: Thanos

Thanos adds a sidecar to Prometheus that uploads blocks to object storage (e.g., S3, GCS) and provides a global query layer across multiple Prometheus instances.

What does the queue_config section within a remote_write block control?

Answer: Tuning parameters for the in-memory send queue such as capacity, batch size, and retry behavior

queue_config lets operators tune how many samples are buffered, how large each batch is, and how Prometheus handles backpressure or failures when writing to remote storage.

Which long-term Prometheus storage solution is developed and maintained by Grafana Labs?

Answer: Mimir

Grafana Mimir is the current open-source, horizontally scalable long-term storage solution for Prometheus developed by Grafana Labs, succeeding Cortex.

What does setting 'honor_labels: true' in a Prometheus scrape_config do when used for federation?

Answer: Preserves labels from the scraped target when they conflict with externally-set labels

honor_labels: true instructs Prometheus to keep the target's own labels intact instead of overwriting them with external labels like job or instance defined in the scrape config.

What flag must be passed to a Prometheus server to allow it to accept remote_write data from other Prometheus instances?

Answer: --web.enable-remote-write-receiver

The --web.enable-remote-write-receiver flag activates the /api/v1/write endpoint so that Prometheus can act as a remote_write target for other Prometheus servers.