#Extending historical NFS connected‑client visibility (>48h) using NetApp Harvest — has anyone done t

1 messages · Page 1 of 1 (latest)

frigid mulch
#

Hi all,
I know the ~48h limit comes from ONTAP’s NFS connected‑clients cache (not Harvest), and that
/api/protocols/nfs/connected-client-settings lets you increase client_retention_interval.
Curious if anyone has actually done this in prod:

What retention did you set?
Any cluster impact (perf/memory)?
Any Harvest or Prometheus tuning needed?
Any gotchas (client unmounts, failovers, cardinality)?

Goal is historical NFS client usage (cleanup/audit), not just real‑time.
Would love to hear real‑world experiences. 🙏

modern prairie
#

if this data is in prometheus, you should have as far back as you need (to whatever you have set your data retention too) to look at this data.

we have a 90 day retention in prom

surreal trail
#

@frigid mulch Please check in the #1062049169520476220 group whether increasing client_retention_interval has any impact on the cluster. You may not need to do that as old data will be stored in prometheus anyways as @modern prairie suggested.

This template is disabled by default in Harvest because it can collect a lot of data. it collects client_ip, which may change frequently. As a result, it’s not a good fit for Prometheus due to high cardinality.