Understanding Prometheus Counter Metrics ๐Ÿ“Š

Understanding Prometheus Counter Metrics ๐Ÿ“Š


3 min read


In the world of monitoring and observability, Prometheus is a powerful tool for collecting and analyzing metrics from your applications and infrastructure. Among the various metric types that Prometheus supports, counters are fundamental for tracking cumulative, always-increasing values. In this blog, we'll delve into Prometheus counter metrics, their usage, and real-world applications, all sprinkled with relevant examples and emojis. ๐Ÿš€

๐Ÿ“ˆ Counters: Metrics That Only Go Up

Counters are like your personal tally, continuously increasing with every occurrence of a particular event. They are perfect for tracking things that naturally accumulate, such as the number of API requests, errors, orders, or bytes sent over a network interface.

Example: Let's say we're measuring the number of API calls for an "add_product" operation:

# HELP http_requests_total Total number of http api requests
# TYPE http_requests_total counter
http_requests_total{api="add_product"} 4633433

The metric name is http_requests_total, labeled with api="add_product", and the counter's value is 4,633,433. This tells us that the "add_product" API has been called 4,633,433 times since the last service start or counter reset.

๐Ÿ’ก Real-Time Use Case 1: API Request Rate

Now, simply knowing the absolute number doesn't provide much insight. However, when used with PromQL's rate function, you can calculate the requests per second the API is receiving.


This query computes the average requests per second over the last five minutes.

๐Ÿ’ก Real-Time Use Case 2: Absolute Change

To calculate the absolute change over a time period, you can use the increase function.


This returns the total number of requests made in the last five minutes. It's essentially the same as multiplying the per-second rate by the number of seconds in the interval (five minutes in our case).

rate(http_requests_total{api="add_product"}[5m]) * 5 * 60

๐ŸŒ More Use Cases for Counters

  1. E-commerce Orders: Counters are perfect for measuring the number of orders in an e-commerce site. As each order is placed, the counter increases, providing real-time statistics on order volume.

  2. Network Traffic: They're great for tracking the number of bytes sent and received over a network interface. These counters help monitor bandwidth utilization.

  3. Application Error Count: To monitor application errors, counters can tally the number of errors, giving a straightforward indication of the error rate.

๐Ÿ Example: Creating and Incrementing a Counter in Python

Using the Prometheus client library for Python, you can easily create and increment a counter metric. For instance, in Python:

from prometheus_client import Counter

api_requests_counter = Counter(
    'Total number of http api requests',

๐Ÿ” Handling Counter Resets

Since counters can be reset to zero, it's important to ensure that the backend used for storing and querying metrics can handle this scenario while still providing accurate results in the event of a counter restart.


Counters in Prometheus are your go-to metric type for tracking cumulative, always-increasing values. They are indispensable for monitoring everything from API requests and network traffic to order volumes and error rates. By understanding how to use counters effectively, you can gain valuable insights into the growth and behavior of your systems. ๐Ÿ“ˆ๐Ÿ”ข๐Ÿš€

Did you find this article valuable?

Support Prasad Suman Mohan by becoming a sponsor. Any amount is appreciated!