This message was deleted.

are you hitting a bandwidth license limit maybe? I saw similar when I had that happen and it started dropping packets

Whilst this was going on I had no notification of the VPX200 limit being reached. I do occasionally receive plateformRateLimitPktDrop in the event log, but that’s very infrequent.

I never saw anything in the logs to indicate it thought it hit a limit but it dropped packets anyway

Okay, the number of connections concerned me.

Citrix told me that the TCP server connections counter is broken in 12.1 and it's just a display issue. I see the same thing on 13.0 so I assumed it was still just a display broken issue. That being said, I do occasionally see high latency spikes here and there, but couldn't attribute it to the ADCs because it has never been a global thing.

I think that might be a bug..

If I look at the system after a failover event the number of connections is down to just over 1000.

yeah, the tcp display thing has been broken for a while

There are no tx drops. The rx drops range from 1-15 (Rate /s), in the past 3 hours there’s been roughly 7,000.

do you have nagles enabled?

Not that’s not in use

I’m out of ideas

Its very odd, the icmp latency usually averages around 12ms from my connection at home and the it spikes from anywhere from 20-200ms when the issue occurs.

Couple of things if your not seeing much on the netscaler in the way of issues dont forget the anciliary stuff outside, can you get the networks team to test firewalls, routers etc to make sure nothing is choking there. maybe the failover GARP is clearing the issu. Also ICMP ping is not always a good test of latency, while it suggests a problem most infrastructure will demote it's priority and send it to the back of the queue so if you are seeing 200ms latency on ICMP does not mean that actual traffic will see the same response time.

Definitely no issues upstream as services going through the same firewall which aren’t load balanced by the ADC have no issues.