Measure latency of Citrix traffic
Not sure where to post this. Citrix Metaframe XPa on WIN2K server. We have occasional slowdowns during Citrix sessions across a WAN. The users complain that sometimes the screen refreshes fall way behind(a full line of text) their typing.
We are measuring bandwidth utilization and network latency and are not seeing any statistics that show us why the users experience these slowdowns. In addition, the servers themselves don't seem to be low on memory or experiencing particularly high utilization at the time of the slowdown.
What I'm looking for is some kind of measurement tool that would at least show us statistically when these slowdowns occur so that we can get a closer look at the problem. If there was a tool that sent traffic over port 1494 and would be able to tell us real time when the slow downs are occuring, we would at least be able to know exactly when the problem is occuring. As it stands now, we're relying on the users to tell us when their experiencing the problem and it makes it difficult to isolate the problem.
Any thoughts on how better to isolate the problem? We're monitoring network latency(ping tool), server utilization, memory usage and memory pages.
We are measuring bandwidth utilization and network latency and are not seeing any statistics that show us why the users experience these slowdowns. In addition, the servers themselves don't seem to be low on memory or experiencing particularly high utilization at the time of the slowdown.
What I'm looking for is some kind of measurement tool that would at least show us statistically when these slowdowns occur so that we can get a closer look at the problem. If there was a tool that sent traffic over port 1494 and would be able to tell us real time when the slow downs are occuring, we would at least be able to know exactly when the problem is occuring. As it stands now, we're relying on the users to tell us when their experiencing the problem and it makes it difficult to isolate the problem.
Any thoughts on how better to isolate the problem? We're monitoring network latency(ping tool), server utilization, memory usage and memory pages.
AvonWyss
How are the users connected to the server? Note that the slowdown may very well e on the side of the users. For instance, if they have a local application which transfers data over the same link (for instance, an email client), they may experience a slowdown while their connection is "jammed up" with non-citrix traffic.
ASKER
Yes and we've had varying degrees of success enabling/disabling/adjusti
From what you say, I don't suppose that it is a limitating factor on your end. Could it be that one of the WAN links is slow during this time?
Are the local clients caching/queuing keystrokes?
ASKER
AvonWyss - Yes. But we have bandwidth monitors running that say we're not maxed and ping utilities indicating that latency is not high.
As for queuing keyboard and mouse, no, but this is true whether there are problems or not.
I guess what I'm looking for is a tool that can emulate and monitor this performance so I can "catch" the problem and have a better chance of identifying it.
As for queuing keyboard and mouse, no, but this is true whether there are problems or not.
I guess what I'm looking for is a tool that can emulate and monitor this performance so I can "catch" the problem and have a better chance of identifying it.
How many users are concurrently connected?
Do you have Citrix Management Services installed on the servers in your server farm?
Have you tried setting up performance monitor counters for CPU activity on the servers in your farm to see if they are spiking at specific times of the day. If these could be tallied to occasions where users report poor response from applications - you might find that you have a misbehaving app.
Have you tried setting up performance monitor counters for CPU activity on the servers in your farm to see if they are spiking at specific times of the day. If these could be tallied to occasions where users report poor response from applications - you might find that you have a misbehaving app.
ASKER
hstiles, do you mean Resource Management Services? If so, no, we don't have it. If not, I don't know.
Sorry, my typo - I meant Resource Management.
In that case, try setting up performance monitor alerts for CPU activity above 60% and a CPU/server work queue of greater than 5. This will alert you to instances of a single process hogging one or more CPUs or creating bottlenecks.
Just to get an idea of numbers, how many users roughly, how many servers in the farm and what is the average server spec. Any 16-bit apps?
In that case, try setting up performance monitor alerts for CPU activity above 60% and a CPU/server work queue of greater than 5. This will alert you to instances of a single process hogging one or more CPUs or creating bottlenecks.
Just to get an idea of numbers, how many users roughly, how many servers in the farm and what is the average server spec. Any 16-bit apps?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
If its possible to give partial credit, it should go to hstiles for coming closest to actually answering the question I asked - which was a tool to better monitor Citrix performance....