Prometheus

Prometheus is a system for collecting and storing metrics from your applications and infrastructure.

The easiest mental model:

Prometheus = a database + monitoring system for numbers over time.

Example

Your backend might expose metrics like:

http_requests_total 152340
http_errors_total 342
http_request_duration_seconds 0.18
active_users 57

Prometheus periodically collects these metrics and stores them.

Your application
      │
      │ metrics
      ▼
  Prometheus
      │
      │ query
      ▼
   Grafana
      │
      ▼
   Dashboard

What is a metric?

A metric is a measurable number.

For example:

CPU usage        → 72%
Memory usage     → 4.2 GB
Requests         → 850/sec
Errors           → 12/sec
Response time    → 180 ms

Prometheus is particularly good at storing these time-series data:

10:00 → 120 requests/sec
10:01 → 145 requests/sec
10:02 → 190 requests/sec
10:03 → 850 requests/sec  ← something happened

Prometheus + Grafana

They’re often used together:

              Application
                   │
                   │ metrics
                   ▼
              Prometheus
                   │
                   │ PromQL
                   ▼
                Grafana
                   │
                   ▼
              📊 Dashboard

Prometheus stores and queries the metrics.

Grafana displays them nicely.

For example, Grafana might ask Prometheus:

"What was the HTTP error rate during the last hour?"

using PromQL (Prometheus Query Language).

Prometheus vs OpenSearch

This distinction is important:

	Prometheus	OpenSearch
Main data	Metrics	Logs / searchable events
Example	CPU = 72%	"Database connection failed"
Data type	Time series	Documents
Query	PromQL	OpenSearch query
Typical visualization	Grafana	OpenSearch Dashboards / Grafana

So your monitoring stack could look like:

                    Application
                         │
                 OpenTelemetry
                         │
          ┌──────────────┼──────────────┐
          ▼              ▼              ▼
        Logs          Metrics         Traces
          │              │              │
          ▼              ▼              ▼
     OpenSearch      Prometheus      Tempo/Jaeger
          │              │              │
          └──────────────┼──────────────┘
                         ▼
                      Grafana

Mental shortcut:

• Prometheus → “How many? How fast? How often?”

• OpenSearch → “What happened?”

• Tracing → “What happened during this particular request?”

• Grafana → “Show me all of this.”

---

Back-End My-Journey-In-Codeless