Data retention with prometheus | NetApp | Page 1

inner minnow Apr 26, 2023, 1:28 PM

#

Hello,

I have some questions about prometheus in docker-compose mode for harvest.

Which is the default data retention?
Which is the best practice for data retention? (I remember some discussions where prometheus was not the right choice for storing >1 year data)

Thanks!

topaz sigil Apr 26, 2023, 3:45 PM

#

hi @inner minnow Prometheus's default data retention can be found at runtime via http://:9090$prom-ip/status
unless it was changed, the compose workflow should be 15d

#

in terms of long term retention perhaps this is the thread you were thinking of? https://github.com/NetApp/harvest/discussions/619

GitHub

Long-term Retention of Metrics · NetApp harvest · Discussion #619

Our organization is a long-time user of the original Netapp Harvest, which used Graphite as its data store. With Graphite, it was easy to retain metrics for long term (1 year or more), and automati...

inner minnow Apr 26, 2023, 4:55 PM

#

Currently I set up 90d via prom-stack.yml, I'll read further the above discussion, thanks

vagrant spindle Apr 26, 2023, 8:33 PM

#

To avoid Prometheus from collecting the same counters repeatedly, you can adjust the scrape_interval parameter in this file: https://github.com/NetApp/harvest/blob/main/container/prometheus/prometheus.yml#L2. It is recommended to increase the duration to be less than a minute (since the performance counter poll schedule is one minute). This way, you can reduce the frequency of data scraping and improve the efficiency of Prometheus.

inner minnow May 5, 2023, 2:34 PM

#

@vagrant spindle you mean to increase scrape_interval, or decrease? Default is 1m if I'm right

#Data retention with prometheus