Grafana / Prometheus DS: no data displayed on small aggregation interval

[smiyc]

When using the new Prometheus based dashboards and selecting a small aggregation interval like 1m, some graphs are not displayed.

To Reproduce
Steps to reproduce the behavior:

1. Go to Database Overview
2. Click on agg_interval
3. Select 1m
4. See that not all graphs are displayed

Expected behavior
see all graphs

Screenshots

with 5m agg_interval

Desktop (please complete the following information):

• OS: windows11
• Browser firefox
• Version 142

Additional context
Add any other context about the problem here.

[pashagolub]

The answer is simple: do not use small aggregation interval. That's how Prometheus scrapper works by design.

Or decrease metric interval + decrease scrapping interval

[smiyc]

thx for the fast reply

the prometheus scrape interval is 15s.
i tried to update the metric interval from the pgwatch gui but face an error

Where is the config persisted in case of an installation on k8s using a yaml file for moitored dbs.
Is it possible to change the interval in such environment?

[smiyc]

got it /pgwatch/metrics/metrics.yaml
so in order to persist it, I'll try to map it as a configMap/volumeMounts to the pod

[pashagolub]

Yeah, you can have your own version of yaml provided as --mertics=path/to/my_metrics.yaml (https://pgwat.ch/v4.x/reference/cli_env.html#metrics)

But you don't have to! You can specify metric by metric for your monitored database with appropriate timeouts, example

[pashagolub]

Also, please remember decreasing metric timeouts will produce a load on a database and on a storage.

So think twice, if you really need short aggregation intervals on the dashboard and what benefit (if at all) it gives you.

[smiyc]

thx