<@ULL7R5KT6> Happy to look into this a little more...
# citrix-vad
s
@Arthur Happy to look into this a little more but I am having a hard time even locating that policy setting mentioned in the post on any supported versions of CVAD. I don't believe it exists.
To the best of my knowledge, there is no concept of session reliability throughput. CGP doesn't throttle at all. It basically retransmits buffers if they are lost.
a
Let me try to test it and will back to you. Thanks
I also can't find that specific policy and I am currently checking this with the author of the post. Here is what I tested in the meantime: Applied the Overall session bandwidth limit set to 20 Kbps. Result: Everything is extremely slow. My observations in Director: a. Output Bandwidth: Shows Available Bandwidth 140 Mbps, Consumed 20 Kbps. There is no alert indicating that a 20 Kbps limit is being hit. b. In-session indicator: It shows "Your connection is excellent", which is highly misleading. The latency inside the session also shows as Excellent (16). c. Report data: After downloading the report, the max latency is 125 ms and the average is around 30 ms, which does not match the ICA RTT I am seeing in Director. d. Latency in Director from 24 ms to 3611 ms It's worth to test it…
b
Reproduced Arthur’s setting. RTT spiked to +1200ms which makes total sense. I’m thinking about how this kind of misconfiguration could be addressed. I’ll probably take this to the Citrix Advisor team.
💯 2
I couldn’t find the respective registry key (the one for the LinkedIn autor) in our internal documentation.
👀 1
a
Here are the concise feature requests for Citrix engineering, based on the observed behavior: 1. Citrix Workspace App (CQI) Improvements ◦ Policy-Aware CQI:The Connection Quality Indicator (CQI) currently relies on raw network capacity. It must be updated to factor in HDX bandwidth limits. If Consumed Bandwidth hits the Policy Limit, the CQI should reflect the degraded experience. ◦ Accurate Status Messages: Instead of misleadingly showing "Excellent" on a throttled connection, the CQI should display: "*Session quality limited by administrative policy*." ◦ Expose True ICA RTT: In-session statistics should show the actual ICA RTT (which includes queuing/buffer delays caused by throttling), rather than just the physical network latency. 2. Citrix Director Improvements ◦ Policy Bottleneck Alerts: Director UI must explicitly flag when a session is artificially limited. Next to the Consumed (20 Kbps) and Available (140 Mbps) metrics, it should show: Throttled by Policy: [Policy Name]. ◦ Fix Reporting Discrepancies: The exported reports flatten extreme latency spikes (e.g., capping at 125 ms when UI shows 3611 ms). Reports must include 95th and 99th percentile ICA RTT metrics to accurately capture queuing delays and micro-bursts, rather than relying on flawed averages and max caps. ◦ Split Latency Metrics: The general "Latency" metric in Director should be separated into three distinct values to stop masking the root cause: 1. Network Latency (physical network delay). 2. Host/Queueing Delay (buffer time caused by bandwidth caps). 3. Total ICA RTT (end-to-end user experience).
💯 2
o
Citrix folks. Give @Arthur a bounty bonus, 💵💲, you know, like they do when cyber-sec folks find bugs and report them to the vendor 😉
a
Haha, thanks! But honestly, the best bounty is just getting an even better product at the end of the day. 😉🚀
🙌 1
👍 1
d
@Arthur I will be proposing an improvement to account for the admin configured throttling
👍 1