The head of IT reported extreme slowness of Teams in his W2019 PVS shared machines since the beginning of the week. Insights showed high latency and low protocol bandwidth in the historical data but no CPU or memory peeks were seen. It was suggested to look at real-time performance views on his machine, and consider upgrading to the latest ControlUp version 8.8 to check all FSLogix metrics.
Read the entire article here...
CPU Related Training & Support Archives
CPU training and support-related archives from inside the ControlUp Community on Slack.
Alert for CPU Temperature in ControlUp
A user asked how to create an alert based on CPU Temperature. It is not collected by default but can be put in a custom index for Macs by using the output from the 'sudo powermetrics' command.
Read the entire article here...
Read the entire article here...
High User Input Delay Issues: Patch Windows OS and Adjust CPU Priority
The user was asked if the high User Input Delay (UIP) was due to infrastructure issues or the Microsoft counter, and @member recommended patching Windows OS if there is a memory leak in ControlUp agent, as well as comparing session count/CPU usage and reducing CPU priority of non-critical processes. The user thanked and will patch as soon as possible. More information can be found in the article at https://support.controlup.com/docs/microsoft-known-issue-user-input-delay-counter-may-cause-handle-leaks-in-controlup-agent.
Read the entire article here...
Read the entire article here...
Performance Monitor Counters – An Overview
@member shared insight into various Performance Monitor Counters, such as CPU Queue Length, Disk Queue Length, Open Handles, Context Switches, and Peak User Input Delay. These metrics measure capacity, thread waiting to be executed, open file handles, rate of context switches, and maximum delay experienced by user input, respectively. Information on these metrics can be found in Microsoft documents (https://systemcenter.wiki/) and a blog post by ControlUp (https://www.controlup.com/resources/blog/).
Read the entire article here...
Read the entire article here...
Investigating Input Delay Causes using ControlUp
A user wanted to understand the cause of an input delay (measured at 60s), but the CPU said 21%, hover-over didn't show anything, and active processes revealed nothing more than 0%. The suggestion was to look at the process_stops index, but there is no way to add this into an alert for easy identification.
Read the entire article here...
Read the entire article here...
Troubleshooting Video Latency on Non-Persistent VM with ControlUp
A user asked about troubleshooting video latency on a non-persistent VM, and it was suggested to use User Session Details in Solve as the best starting point. If Remote DX was installed, they can also watch session relevant latency, CPU, RAM and other metrics in a single place. Script Action can be used for advanced data collection. The Network Latency graph may be legit, but there was a known issue with the Linux RDX plugin. A new beta version which fixes this issue will be available soon. After sifting through the data, 3.75GB of RAM usage was found - they will look at adjusting graphics settings with Horizon View.
Read the entire article here...
Read the entire article here...
CPU Data Points Not Showing Any Service Processes in ControlUp
A member reported an issue where the data points on the CPU graph do not show any service processes. This is happening on all of Eugene's devices running Windows and macOS. ControlUp's support will check the issue and reach out to Eugene with a support ticket.
Read the entire article here...
Read the entire article here...
Troubleshooting 0% CPU for Edge DX Endpoints
An issue arose where all CPU was 0% for a group of Edge DX endpoints with no data in the process_stops data index. It was postulated that a collection issue was the root cause, and as a first step the user was encouraged to restart the SIP agent and check the application event log.
Read the entire article here...
Read the entire article here...
Stress Testing Tool Recommendations for Generating High Usage in ControlUp
A user asked for recommendations on stress testing tools and generating high CPU/RAM/network usage. Suggestions offered included Notepad for increasing memory, CPUStress, LoadGen, SimLoads from eucscore, a PowerShell script for maxing out all but one of the CPU cores, and a Linux command to generate a single thread of stress. The G0-EUC team use LoadGen for their research and the post also provided a link to their testing methodology.
Read the entire article here...
Read the entire article here...
Automating the Disabling of Maintenance Mode on Virtual Desktops
A user asked whether it was possible to automate the process of disabling maintenance mode on virtual desktops. Suggestions include marking the systems by using the eventlog, using registry keys for non-persistent systems, and using a local database. Citrix's Load Index was mentioned as a simpler solution, and it was clarified that it uses CPU, RAM, and maximum user numbers to calculate the effective load. ControlUp Automation and Citrix Virtual Apps and Desktops Policy Settings were both suggested as useful solutions.
Read the entire article here...
Read the entire article here...