Grafana shows only 15 days of metrics instead of 90 days

Grafana Prometheus
, ,
1

I’m using Grafana to visualize CPU, memory, and disk metrics from Prometheus using Node Exporter as the data source. However, I noticed that my Grafana dashboard only displays historical data for the past 15 days, even though I’ve already configured the data retention period to 90 days in Prometheus.

image
image505×322 4.56 KB

image
image718×542 12.7 KB

Could you please help me understand what might be causing this issue?
Is there a way to verify or troubleshoot whether the data retention settings are actually applied, or if old data is being deleted due to storage limits or other configurations?

2

Welcome @feetee00500 to the Grafana forum.

Which version of Grafana are you using (Cloud or OSS)? If Cloud, are you on a paid tier? The free tier’s storage of Metrics is limited to 14 days, regardless of what your Prometheus settings are.

3

I am using Grafana (OSS) installed in a Docker container on my on-premises system. As you mentioned:

  1. Grafana (OSS) stores data for a maximum of 15 days — is that correct?

  2. Grafana (OSS) can store unlimited data depending on the Prometheus configuration — is that correct?

If I don’t understand, please explain it to me again. If I want to store data for up to 90 days in the past, what should I do?

image
image950×508 11.5 KB

4

Hi @feetee00500 Grafana (Cloud or OSS) does not store any data. You connect it to a datasource of your choice. That datasource can hold 5 minutes or 5 decades of data depending on how you configure it.

It sounds like your Prometheus datasource is also on-prem. By default, Prometheus stores data for 15 days and you can change this by changing the retention period, which is controlled by a command-line flag you set when starting Prometheus.

You can control data retention in Prometheus using two primary flags:

  • Time-Based Retention: Use the --storage.tsdb.retention.time flag to set how long to keep data. You can use units like d for days, w for weeks, and y for years.

  • Size-Based Retention: Use the --storage.tsdb.retention.size flag to set a maximum size for the storage blocks (e.g., 50GB).

If you set both flags, Prometheus will delete data whenever the first limit is reached.