#snapshot data and/or creation dates available?

1 messages · Page 1 of 1 (latest)

crude burrow
#

snapshot show -volume volume1 -fields vserver,volume,snapshot,create-time,size,snapmirror-label

is there a harvest metric that has this information? I am not seeing this on my harvest servers anywhere and the logs are showing this:

time=2026-04-27T13:12:14.485-07:00 level=INFO source=collector.go:415 msg="no instances, entering standby" Poller=vserver7 collector=Rest:Snapshot task=data

#

we have a number of snapshots on this specific cluster

#

snapshot data and/or creation dates available?

crude burrow
#

well, upon reviewing what is there, the snapshot info is basically the newest snapshot and the oldest snapshot for that specific volume. that limits all the records down to 2 per volume instead of many per volume.

#

we are currently creating a separate metric from CLI for tracking the snapshots on our systems (making sure we don't have abandoned snapshots or no snapshot (depending on the case)

halcyon zodiac
crude burrow
halcyon zodiac
#

Thanks

crude burrow
#

Any chance this will make it into 2026.05?

muted plaza
#

hi @crude burrow this feature probably won't make it for 2026.05

halcyon zodiac
#

@crude burrow Could you share more details on how you plan to use the first and last snapshot information for a volume? We’re evaluating whether we can implement the functionality you’re looking for directly in Harvest using this data.

For example, we already have support for snapshot policy compliance (disabled by default), as described here:
https://github.com/NetApp/harvest/issues/3798

crude burrow
#

we need to know the latest snapshot and the oldest snapshot. This allows us to track if there are snapshots that are 'abandoned' or volumes 'no longer producing a snapshot' by having a query designed around these two sets of data. ie. if the oldest snapshot is over 425 days old, flag these snapshots for deletion (we only keep them for up to 425 days). If there are no snapshots in the past 365 days, check as to whether we can delete the old snapmirror (vault) volume (as we only keep vaulted volumes for up to 365 days). This is why we need the newest and oldest snapshot information.

#

hope this helps

halcyon zodiac
#

Thanks @crude burrow for explanation. Are days here in query static or they can be dynamic?

crude burrow
#

my days are a set value that is static, it doesn't change when the snapshots age. Its a required timeline we are expected to hold snapshots for vaults

#

if we are recording the snapshot age, I would expect that information to be static, not changing since whatever the filer holds would be reported as the date of creation.

#

The query for the maximum age of our vault volume snapshots in our dashboard is similar to this:

"expr": "(sum by (volume,aggregate) (time() - snapshot_age{volume=~"$Volume",vserver=~"$Vserver"})) ",

#

we have a alert (via alertmanager) that is triggered via this similar query:

expr: (time()-snapshot_age > 36720000)

#

so when snapshots get to be older than 425 days old this will trigger

#

having just the newest and oldest keeps the number of snapshot records to the lowest number we can without flooding prometheus with data we won't use