Remote Desktop Services (RDS) connectivity is usually reliable, but things can -- and sometimes do -- go wrong.
Typically, user connectivity issues (i.e. slow refresh, sluggish performance, etc) is due to insufficient bandwidth and/or latency in the connection between the user device (computer, laptop, phone, etc) hosting the user Desktop and the server in GCP acting as the RDS Host.
To use a CAS provisioned Cloud Desktop, each user needs to have a provisioned Desktop, the correct login credentials and a broadband internet connection.
User Bandwidth and Latency Requirements
For optimal performance, the local connection should have 250 kilobits per second (kbps) of bandwidth per user. However, 150 kbps per user will suffice.
The key is latency. Latency of under 100 milliseconds (ms) is recommended for a good user experience.
If there are issues with either, please contact your Internet Service Provider (ISP) to ensure your contract/Quality of Service (QoS) meets your needs.
GCP Connectivity: Testing Latency
To measure the user's latency to the GCP Region hosting their RDS Host, use this Google tool: www.gcping.com
If the latency is consistently great than 100 ms, then resolution might mean moving the customer to a closer GCP Region.
User Behaviour and Connectivity
The quality of the user connection isn't just a by-product of bandwidth or latency, its also a function of what the user is doing in their Cloud Workspace. Here's a set of questions to ask:
What applications is the user running?
How large are the files being opened by the user?
Is the customer VM (RDS Host) properly provisioned?
In other words, if the user is opening too many (resource intensive) applications while working with many large files on a customer VM (RDS Host) which is under-provisioning (i.e. not enough vCPU, RAM or Disk Space), then the user experience will suffer.
The same applies if the customer VM (RDS Host) is hosting too many users.
Either situation can be resolve quickly in CAS by increasing resource allocations on the customer VM, using Dynamic Resourcing or adding a second RDS Host (VM) to distribute the load.
Google Chrome Tabs
Some users prefer to have 20, 30 or more tabs open in Google Chrome. There will definitely be performance issues because of known issue with Chrome consuming too much memory when too many tabs are open.
Since its often difficult to change user behaviour, we recommend implementing 3rd party tools such as Tab Wrangler to automatically close unused tabs to free-up VM memory.
Remote Management and Monitoring (RMM) Tools
The key to being proactive, as well as having critical info/real-time data available, is implementing an RMM tool. Remote Management and Monitoring (RMM) tools provide Dashboard-level data and reports on each customer VM against a wide-variety of criteria: monitor and track CPU & RAM usages, proactively detect and resolve threats, help with automation, etc.
itopia recommends SolarWinds. a enterprise class IT management and monitoring tool.
Advanced Troubleshooting
The Windows Event Viewer provides objective data on what exactly is happening on the customer's VM (RDS Host). Essentially, the Windows Event Viewer allows Windows admins to 'retrace the steps' of the user to see where exactly the problem might be.
Windows Event Codes
Two examples with links to Microsoft Help
Unexpected Shutdown: 41
Service didn't start promptly: 7000 & 7011
Related Articles
Update VM Instance Resources (vCPU and RAM)