Skip to Main Content
DoiT logo
Pillar Optimization
Categories Lenses
Created by Guest
Created on Jul 22, 2022

GCP Lens show confusing CUD Coverage

Thanks, Andrew.

I see your meaning, but the graphs of GCP Lens are misleading. For example, if CUD is 10 and only being used 5, then GCP Lens will show 200% here, instead of 50% which is commonly understood. Hope my feedback could lead to some user-friendly changes.

Thanks,FrankFrom: Andrew Lee <>
Sent: Thursday, July 21, 2022 1:00 AM
To: Frank Li <>
Cc: Daniel Anderson <>; Shawn Fang <>; Eric Gao <>; Ronger Chen <>; Ben Yu <>
Subject: Re: GCP Lens show a big spike on CPU/RAM

Caution: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

Hi Frank,

Thank you for reaching out about this. The chart that you referenced is essentially saying that you currently have more CUDs than you are using, which means your CUDs for those regions are being underutilized. Below is an area chart showing this. The flat graph is your CUDs you currently have in us-west2, while the red graph is your usage. As you can see, starting around 7/12 to 7/13 that usage drops sharply, causing the chart you referenced to display that you are over-covered here by CUDs.

Here is a link to the report as a reference:

Was this change intended? Is it possible to move more workloads to this region to minimize underutilization? In looking at your numbers currently, it seems like you have CUD coverage of 84 cores as you mentioned, but only 12 cores being used in that region.

Let me know if you have any other questions on this.

Thank you,

Andrew Lee

Technical Account Manager
2022 PTO and Holiday Calendar: 9/5, 9/16 - 9/23, 10/10, 11/24 - 11/25, 12/26

425 900 1390

We're Hiring! Check out our open positions and see what it's like to work at DoiT.

On Tue, Jul 19, 2022 at 8:24 PM Frank Li <> wrote:

Hi Andrew/Daniel,

As Shawn invites me to join the DoiT CMP, I find GCP usage has a big increase on CPU/RAM recently.

The top line is us-west2, I don't understand what 700% means here. Since we do have GCP CUD on N1 in us-west2 as 21 instances (84 cores, 126GB ram), and we are running only 11 instances now. Then I checked the second top line northamerica-northeast1 (200%), same issue too.

Could you please explain how it comes?


This e-mail message and all attac

  • Attach files