Abhishek Kumar

<p>Modern systems generate tons of metrics — CPU usage, memory consumption, request rates, error counts, latency, and much more. To monitor and analyze all of this efficiently, developers and DevOps engineers often rely on <strong>Prometheus</strong>, one of the most popular open-source monitoring and alerting tools.</p><p>In this blog, we’ll introduce Prometheus, set it up quickly, and walk through <strong>PromQL</strong>, the query language that makes Prometheus powerful.</p><h2>What is Prometheus?</h2><p>Prometheus is an open-source monitoring and alerting system originally built at SoundCloud. It has become the de facto standard for monitoring <strong>cloud-native environments</strong> like Kubernetes.</p><p><strong>Key features of Prometheus:</strong></p><ul><li><p><strong>Time-series database</strong> – stores metrics as time-stamped data.</p></li><li><p><strong>Pull model</strong> – Prometheus scrapes (pulls) metrics from exporters or instrumented apps.</p></li><li><p><strong>Powerful query language (PromQL)</strong> – allows flexible analysis of metrics.</p></li><li><p><strong>Alertmanager</strong> – for handling alerts based on rules.</p></li><li><p><strong>Rich ecosystem</strong> – exporters exist for databases, OS metrics, Kubernetes, and more.</p></li></ul><h2>Getting Started with Prometheus</h2><h3>Step 1: Install Prometheus</h3><p>You can download Prometheus directly:</p><pre><code class="language-bash">wget https://github.com/prometheus/prometheus/releases/download/v2.54.1/prometheus-2.54.1.linux-amd64.tar.gz
tar xvfz prometheus-*.tar.gz
cd prometheus-*/</code></pre><h3>Step 2: Run Prometheus</h3><pre><code class="language-bash">./prometheus --config.file=prometheus.yml</code></pre><p>By default, it starts on port <strong>9090</strong>. Open <a target="_blank" rel="noopener noreferrer nofollow" href="http://localhost:9090">http://localhost:9090</a> to see the UI.</p><h3>Step 3: Configure Targets</h3><p>In <code>prometheus.yml</code>, you can define scraping targets:</p><pre><code class="language-yaml">scrape_configs:
  - job_name: 'node'
    static_configs:
      - targets: ['localhost:9100']</code></pre><p>Here, Prometheus scrapes metrics from <strong>Node Exporter</strong> running on port <code>9100</code>.</p><h2>Introduction to PromQL</h2><p>PromQL (Prometheus Query Language) is what makes Prometheus special. It allows you to <strong>query, aggregate, and visualize metrics</strong>.</p><h3>Basic PromQL Examples</h3><ol><li><p><strong>Instant Vector – Current CPU usage</strong></p></li></ol><pre><code class="language-promql">node_cpu_seconds_total</code></pre><p>Shows the raw CPU seconds counter.</p><ol start="2"><li><p><strong>Rate of CPU usage per second</strong></p></li></ol><pre><code class="language-promql">rate(node_cpu_seconds_total[5m])</code></pre><p>Calculates the per-second average increase over the last 5 minutes.</p><ol start="3"><li><p><strong>Average CPU usage across all cores</strong></p></li></ol><pre><code class="language-promql">avg(rate(node_cpu_seconds_total[5m])) by (instance)</code></pre><ol start="4"><li><p><strong>Memory Usage</strong></p></li></ol><pre><code class="language-promql">node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes</code></pre><p>Gives memory utilization ratio.</p><ol start="5"><li><p><strong>HTTP Request Rate</strong></p></li></ol><pre><code class="language-promql">rate(http_requests_total[1m])
</code></pre><p>Requests per second averaged over 1 minute.</p><h2>Visualizing Metrics</h2><p>Prometheus comes with a basic graph UI, but it’s often paired with <strong>Grafana</strong> to create rich dashboards. Using PromQL queries inside Grafana, you can build interactive monitoring dashboards.</p><h2>Alerts with Prometheus</h2><p>Example alert rule (in <code>alert.rules.yml</code>):</p><pre><code class="language-yaml">groups:
- name: example
  rules:
  - alert: HighCPUUsage
    expr: avg(rate(node_cpu_seconds_total[5m])) &gt; 0.9
    for: 2m
    labels:
      severity: warning
    annotations:
      summary: "CPU usage is very high"
</code></pre><p>This triggers an alert if CPU usage is over 90% for more than 2 minutes.</p><h2>Conclusion</h2><p>Prometheus provides a <strong>powerful and flexible monitoring solution</strong> for modern infrastructure. Its pull-based architecture, time-series database, and PromQL make it ideal for monitoring microservices, Kubernetes clusters, or traditional servers.</p><p>By mastering <strong>PromQL</strong>, you unlock the real power of Prometheus, enabling you to query, visualize, and alert on any metric your systems expose.</p>

PromQL

Introduction to Prometheus and PromQL