-
Notifications
You must be signed in to change notification settings - Fork 4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
CA: Update AWS instance type list #2550
CA: Update AWS instance type list #2550
Conversation
Mostly important due to existing Memory figures for some C5d instances being incorrect
15606c7
to
6bceab7
Compare
/assign @Jeffwan |
@@ -223,7 +223,7 @@ var InstanceTypes = map[string]*InstanceType{ | |||
"c5d.12xlarge": { | |||
InstanceType: "c5d.12xlarge", | |||
VCPU: 48, | |||
MemoryMb: 131072, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Em..Interesting. I do use the generator to populate the specs. Do you change manually or use the gen.go
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah, I was confused as well when I noticed it, even weirder I changed it by running make generate
again, so I can only assume the API was returning different results when you last generated it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah. The new values you give match with values in aws docs. Great!
"u-18tb1": { | ||
InstanceType: "u-18tb1", | ||
VCPU: 448, | ||
MemoryMb: 0, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Any reason it get 0 MemoryMb?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ah, I failed to notice that! I think I'll need to dig into this and figure out why that's happening.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Finally got around to spending more time on digging into this, and it looks like the Pricing API we're currently using can't be fully trusted. For one thing the correct API designation of these instances is u-18tb1.metal
, not u-18tb1
as returned by the pricing API.
I'm going to raise a separate issue to discuss whether/how we move away from the Pricing API as it seems it can't be fully trusted to build a proper picture of these nodes, in the meantime any problem with merging this as is? (I've updated it with a couple more changes generated when running today and double checked them against the info returned by instance descriptions using a boto3 client.)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Em.. u-18tb1.metal
is the right name according to release note. I don't have concern merging this since this instances are available from dedicated hosts with 3 yrs reservation. Most users may not even have access to it. Feel free to open a new issue and let's resolve the pricing api issue there. Some other projects uses nodes template as well like CNI, I will also have a check on reliable API.
d5cddf7
to
9211366
Compare
/lgtm |
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: Jeffwan The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
Mostly important due to existing Memory figures for some C5d instances being incorrect.
Despite the ability for the CA to dynamically generate this list added by #2240 users can still use this list and the memory figures for some of the C5d family currently are incorrect. These new generated figures match the published figures: https://aws.amazon.com/ec2/instance-types/c5/
This also pulls in the figures for the new super high memory instances: https://aws.amazon.com/blogs/aws/ec2-high-memory-update-new-18-tb-and-24-tb-instances/