How to Implement Moving Averages in Grafana Dashboards

To add moving averages to your Grafana dashboards using popular data sources like Prometheus and InfluxDB.

Begin by accessing your Grafana dashboard. To incorporate a moving average into your visualization, click the Menu button and select Edit to modify your current dashboard.

Ensure you choose the correct data source from the dropdown menu. Prometheus and InfluxDB are widely favored for implementing moving averages due to their robust support for time-series data.

Moving Averages with Prometheus

Prometheus offers the avg_over_time() function, which effectively serves the purpose of calculating moving averages. While it doesn't have a dedicated moving_average() function, avg_over_time() allows you to compute the average of a metric over a specified time window.

For example, to calculate a 5-minute moving average of HTTP requests for the frontend job, you can use the following Prometheus query:

avg_over_time(rate(http_requests_total{job="frontend"}[1m])[5m:])

Prometheus does not natively support more sophisticated moving averages like Weighted Moving Averages (WMA) or Exponential Moving Averages (EMA). Consider leveraging Grafana's transformation features or integrating with InfluxDB for these advanced calculations.

Using the moving_average() Function in InfluxDB

To implement a moving average in InfluxDB, you can leverage the moving_average() function directly within your InfluxQL queries. This function simplifies the process of smoothing out data by calculating the average over a specified number of data points, effectively highlighting long-term trends while minimizing short-term fluctuations.

The basic syntax for the moving_average() function is as follows:

moving_average(metric, windowSize)

Suppose you want to monitor the memory usage of your application servers to identify long-term usage patterns and mitigate short-term spikes. Calculating a 15-point moving average can provide a clearer view of memory usage trends. Here's how you can structure your InfluxQL query:

SELECT moving_average(mean("memory_usage"), 15) 
FROM "system_metrics" 
WHERE time > now() - 3h 
GROUP BY time(10m)

By effectivelyusing the moving_average() function in InfluxDB and adhering to these best practices, you can create insightful and responsive Grafana dashboards.

Got an article suggestion? Let us know

Explore more

How to install Prometheus and Grafana on Kubernetes with Helm

How to Set Up a Custom Dashboard on Grafana's Home Page

How to Uninstall Grafana?

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.