Earlier this week, I was working with a customer who is using Goliath for Citrix Monitoring and Citrix Troubleshooting for Citrix XenDesktop 7.6 applications. This customer is a law firm that has 19 offices located all over the world. In total, they have over 1,000 end user and 300 VMware vSphere virtual machines supporting their Citrix XenDesktop 7.6 environment.
Citrix Troubleshooting Issue: Complaints about Slow Logon Times
The Citrix Troubleshooting engagement was to focus on was reducing end user complaints resulting from application and server issues, specifically with slow logon times. By configuring our product they were able to isolate the root cause of slow logon times and be alerted to impending issues proactively so they will now have time to react and solve problems before end users complain. This blog will speak to how the law firm achieved their goal.
Citrix Troubleshooting Step #1 – Deploy the Intelligent Agent to Collect Data
So the first Citrix Troubleshooting step was to bring in their VMware vSphere and Citrix XenDesktop 7.6 environment through our ‘First Time Run Wizard’ using our APIs for VMware and Citrix. Once the Citrix infrastructure was listed in the inventory, we deployed our Intelligent Agent.
Our agent is not like most agents. It is very lightweight and uses less than 0.1% CPU, 1.5 MB on disk, and about 50 to 80 MB RAM. The agent deployed to the environment within minutes and instantly started to collect data. Once deployed, we were then able to jump right into configuring the product.
Citrix Troubleshooting Step #2 – Analyze Logon Times & Drill Down into Logon Duration
The second Citrix Troubleshooting step was to start analyzing their logon times and drill down into the logon duration breakdown, the customer went to the XenApp/XenDesktop Session Display. From this display, you can view current and historical Citrix sessions and see metrics including logon duration, ICA latency, connection times, and more.
Since the customer primarily wanted us to troubleshoot slow logon duration times, we then sorted this display by clicking the ‘Logon’ header to bring the sessions with the slowest logon times to the top of the screen, as shown in the below screenshot.
The customer was then able to click into the sessions and view 3 levels of Logon Duration information. These levels included the following and are pictured below:
- The high level breakdown of the logon duration for each of the logon stages
- The breakdown of the logon as seen by the broker/controller
- The logon details as seen by the session itself
When the customer saw all of logon data that Goliath received they were blown away and stated that the detail was “amazing” and that they “didn’t know this detail was possible.”
Citrix Troubleshooting Step #3 – Drill Down into Group Policy & Logon Duration Detail to Identify End User Issues
With the ability to drill down into group policy information, they were able to view what their group policies were actually doing. The customer was able to see each of the shared drives being mapped, printer mappings, and group policy extensions being processed.
Using this Logon Duration detail, the customer was able to identify multiple problems in the users’ sessions loading right away. One item included a group policy extension that was taking 27 seconds, which is higher than normal, pictured below. Using this information, the customer was able to identify the group policy object that it came from, open it, and see an entire registry tree that was unnecessary was being added in.
Overall, having visibility into the logon duration process was hugely impactful for this customer. Since implementing our technology, they have been able to identify multiple areas that were causing logon times to increase. They have also implemented alerting and reporting around logon duration as well.
Citrix Troubleshooting Step 4 – Get Proactive by Configuring Preemptive Alerts Based on Thresholds
The final Citrix troubleshooting step was to use the root cause information to be proactive by configuring preemptive alerts based on thresholds we determined from the issue isolation trending analysis. If leading indicators begin to occur, an alert will be triggered so they now have time to proactively address the issue before users are impacted.
It is very rewarding to implement the product and achieve these types of results for customers who have been frustrated in the past. Stay tuned for additional “proactive stories from the IT front lines”.