How and where we are collecting the QoS wait time field in OCI

Last updated
Save as PDF
Share
1. Share
2. Tweet
3. Share

Views:: 220

Visibility:: Public

Votes:: 2

Category:: oncommand-insight

Specialty:: oci

Last Updated:

Applies to

7.3.9 OnCommand insight

Answer

OCI collects QOS wait time using 'workload_queue_nblade' object and extract 'delay_qos_limit_wait_time' from it.

The request looks like this:

<<< REQUEST >>>

perf-object-get-instances|||objectname=workload_queue_nblade,instance-uuids=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx;xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx

workload_queue_nblade

delay_qos_limit_wait_time

xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx

<<< REQUESTEND >>>

The response back is this:

Counter info:

The raw QoS Limit wait time. Units are microseconds.

false

delay_qos_limit_wait_time

diag

delta,no-zero-values

microsec

Additional Information

So this is how OCI is seeing that as a wait time. it's time spent in waiting due to QoS policies, so it should represent delay in response (possibly due to QoS throttling).

For NetApp Cmode, it's polled from counter "delay_qos_limit_wait_time", for workload_queue_nblade object on internal volume

OCI does convert the unit microsec to millisec in the back ground for easier read for end user.

Depending on how OCI is set for polling time depends on what the QoS Wait Time will show, it will avg out total time collected.