#QOS QTree Stats Anomaly

1 messages · Page 1 of 1 (latest)

nova kernel
#

Hi Gurus
I looking for some assitance understanding some metrics from Harvest which don't appear to add up.
For example, we get the following qos_read_latency of 0, qos_write_latency of 0, but qos_latency of 1206 for a qtree:

qos_read_latency{datacenter="XXXX",cluster="XXXX",wid="XXXX",svm="XXXX",volume="XXXX",qtree="XXXX",lun="",file="",policy_group="XXXX_qos_policy"} 0
qos_latency{datacenter="XXXX",cluster="XXXX",wid="XXXX",svm="XXXX",volume="XXXX",qtree="XXXX",lun="",file="",policy_group="XXXX_qos_policy"} 1206.714069591528
qos_write_latency{datacenter="XXXX",cluster="XXXX",wid="XXXX",svm="XXXX",volume="XXXX",qtree="XXXX",lun="",file="",policy_group="XXXX_qos_policy"} 0

Env info:

  • Harvest: 22.11.1
  • CDOT: 9.8

Thanks in advance!

hasty flume
#

@nova kernel Latency counter in qos is not equal to sum of read and write latency. ONTAP workload latency counter measure the latency of specific types of I/O operations, such as read, write, and metadata operations. However, other factors, such as network latency, queueing delays, and other overhead, can also contribute to overall latency.

That being said, Could you share logs from your system. We'll take a look to see if there are any issues to rule out any Harvest-related problems.

nova kernel
#

@hasty flume Thanks for the informative response! I believe that explains the issue but I am checking to see if it's acceptable to pull the logs for further analysis. Will let you know. Thanks again!!

hasty flume
nova kernel
#

Hi @hasty flume ! Looks like we're not going to upload the logs at this time. Many thanks for your help!