#####   ######  #####    ###    #   #  ###  #   #  ######
##  ##  ##      ##  ##  ## ##   #   #   #   #   #  ##
#####   ####    #####   #   #   #   #   #   #   #  ####
##  #   ##      ##      ## ##    # #    #    # #   ##
##   #  ######  ##       ###      #    ###    #    ######

$ curl repovive.com/roadmaps/devops-sre-interview-prep/observability-and-monitoring/prometheus-and-metrics

░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░██████████████████████████████████████████████████████████████████████████████████████████████████████████

#####   ######  #####    ###    #   #  ###  #   #  ######
##  ##  ##      ##  ##  ## ##   #   #   #   #   #  ##
#####   ####    #####   #   #   #   #   #   #   #  ####
##  #   ##      ##      ## ##    # #    #    # #   ##
##   #  ######  ##       ###      #    ###    #    ######

$ curl repovive.com/roadmaps/devops-sre-interview-prep/observability-and-monitoring/prometheus-and-metrics

Repovive

Prometheus and Metrics - Observability and Monitoring | DevOps/SRE Interview Prep | Repovive

Repovive.

DevOps/SRE Interview Prep16 sections · 139 units8h total

Open in Course

Prometheus and Metrics

Time-series data collection

Prometheus is the standard for metrics in cloud-native environments.

How it works:

Pull-based: Prometheus scrapes metrics from targets
Time-series database stores metrics with labels
PromQL for querying

Metric types:

Counter: Only increases (requests_total)
Gauge: Can go up or down (temperature)
Histogram: Distribution of values (request_duration)
Summary: Calculates percentiles client-side. Cannot aggregate across instances

Example PromQL:

rate(http_requests_total[5m])  # Requests per second
histogram_quantile(0.99, rate(request_duration_bucket[5m]))