# When are summarizers (e.g. 'Total') in a legend useful?

**URL:** <https://community.grafana.com/t/when-are-summarizers-e-g-total-in-a-legend-useful/153651>\
**Category:** Time Series Panel\
**Tags:** legend\
**Created:** [June 30, 2025, 3:42pm UTC](https://community.grafana.com/t/when-are-summarizers-e-g-total-in-a-legend-useful/153651 "2025-06-30T15:42:26Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![vanilladonut](https://avatars.discourse-cdn.com/v4/letter/v/eb8c5e/32.png) [@vanilladonut](https://community.grafana.com/u/vanilladonut)\
**Post date:** [June 30, 2025, 3:42pm UTC](https://community.grafana.com/t/when-are-summarizers-e-g-total-in-a-legend-useful/153651/1 "2025-06-30T15:42:26Z")

</div>

- What Grafana version and what operating system are you using?

- **What** are you trying to achieve?  
Truly understand the meaning of ‘Total’ and other values shown in a Legend.

- **How** are you trying to achieve it?  
By perusing the web, esp. this article. Relevant extract:

> Let us now focus on the first two points. Remember, each dot represents the increment in requests in the time between the dot and 1 minute behind it. If we have 20 points in 5 minutes, between two points, we have an interval of ~5\*60/(20-1)= ~16sec. Now consider the following scheme:

```auto
                     P1 P2
|---------60s--------|--16s--|
|--16s--|---------60s--------|

```

Is Total really overcounting, as implied by the P1 and P2 overlap above?

- What happened?  
I don’t know enough about Grafana to trust the linked article with certainty, although [related web results](https://www.google.com/search?q=grafana+When+are+summarizers+%28e.g.+%27Total%27%29+in+a+legend+useful%3F) do seem to support it. For [example](https://github.com/grafana/grafana/issues/627#issuecomment-50504285):

> Total in legend value is always a sum of the points returned from InfluxDB.

My datasource is Prometheus, though I suspect the same holds true for it.

The opening of the article’s last section probably summarizes its position:

> Well, maybe the best option is not to use any summarizer in the series legend.

Is the article’s position valid, **there really is no legitimate use case for “any summarizer in the series legend”?**