MX100 - Has anyone noticed a higher Device utilization ?

[thomasthomsen]

A customer of mine upgraded from , and Im guessing here, because the change log does not tell us what software you where running before, 15.x (fairly new) to 15.42.3 most likely because the system was forcing /telling him to, and then again to 15.44 (most likely again, because the system was forcing him to, or told him so).

 After this, his utilization spiked.

1: first update during the night

2: second upgrade during the night (not much change here in utilization).

Have anyone else seen something like this ?

I am also chasing other reasons why the utilization has spiked, but it just fits the software upgrade.

(no change have been made to the configuration)

/Thomas

[thomasthomsen]

Just the same graph, without me drawing on it.

and the normal network graph for the same days.

There does not seem to be a higher amount of traffic passing through.

[DarrenOC]

Hi @thomasthomsen , we’ve seen this before. If you roll back you should see the cpu drop as well.

[PhilipDAth]

I have had people tell me about this as well.  I believe some of the 15.x trains simply require more CPU.  I don't see this changing going forward.  It's a natural evolution of adding more features.

[thomasthomsen]

But between two versions in the same release train there should not be that many "new features".

Sure I understand if you upgraded to 16.x or 17.x when it arrives, but within 15.x and no new configuration done on the box, that just indicates to me that something else is going on.

[Owen]

Yes we have also noticed higher CPU and lower VPN throughput as a result of this issue.

There is a note in 16.4 which states "Significant performance regressions for VPN traffic may be observed on MX84 and MX100 appliances" and I suspect this is present prior to this release.