WingNews logo WingNews
top | item 29474172

(no title)

temuze | 4 years ago

Friends tell friends to pick us-east-2.

Virginia is for lovers, Ohio is for availability.

discuss

order

blahyawnblah|4 years ago

Lots of services are only in us-east-1. The sso system isn't working 100% right now so that's where I assume it's hosted.

skwirl|4 years ago

Yeah, there are "global" services which are actually secretly us-east-1 services as that is the region they use for internal data storage and orchestration. I can't launch instances with OpsWorks (not a very widely used service, I'd imagine) even if those instances are in stacks outside of us-east-1. I suspect Route53 and CloudFront will also have issues.

johnsimer|4 years ago

Yeah I can't log in with our external SAML SSO to our AWS dashboard to manage our us-east-2 resources. . . . Because our auth is apparently routed thru us-east-1 STS

jhugo|4 years ago

You can pick the region for SSO — or even use multiple. Ours is in ap-southeast-1 and working fine — but then the console that it signs us into is only partially working presumably due to dependencies on us-east-1.

sbr464|4 years ago

Ohio's actual motto funnily kind of fits here:

  With God, all things are possible

shepherdjerred|4 years ago

Does this imply Virginia is Godless?

more_corn|4 years ago

This is funny, but true. I've been avoiding us-east-1 simply because thats where everyone else is. Spot instances are also less likely to be expensive in less utilized regions.

api|4 years ago

I live in Ohio and can confirm. If the Earth were destroyed by an asteroid Ohio would be left floating out there somehow holding onto an atmosphere for about ten years.

stephenr|4 years ago

Friends tell Friends not to use Rube Goldberg machines as their infrastructure layer.

tgtweak|4 years ago

If you're not multi-cloud in 2021 and are expecting 5-9's, I feel bad for you.

temuze|4 years ago

If you're having SLA problems I feel bad for you son

I got two 9 problems cuz of us-east-1

unethical_ban|4 years ago

If you're not multi-region, I feel bad for you.

If your company is shoehorning you into using multiple clouds and learning a dozen products, IAM and CICD dialects simultaneously because "being cloud dependent is bad", I feel bad for you.

Doing one cloud correctly from a current DevSecOps perspective is a multi-year ask. I estimate it takes about 25 people working full time on managing and securing infrastructure per cloud, minimum. This does not include certain matrixed people from legacy network/IAM teams. If you have the people, go for it.

post-it|4 years ago

I imagine there are very few businesses where the extra cost of going multi-cloud is smaller than the cost of being down during AWS outages.

ryuta|4 years ago

How do you become multi-cloud if your root domain is in Route53? Have Backup domains on the client side?

bithavoc|4 years ago

Sometimes you can't avoid us-east-1; an example is AWS ECR Public. It's a shame. Meanwhile, DockerHub is up and running even when it's in EC2 itself.

mountainofdeath|4 years ago

us-east-1 is a cursed region. It's old, full of one-off patches to work and tends to be the first big region released to.

vrocmod|4 years ago

No one was in the room where it happened

kavok|4 years ago

Didn't us-east-2 have an issue last week?

politician|4 years ago

Can I get that on a license plate?

PopeUrbanX|4 years ago

I wonder why AWS has Ohio and Virginia but no region in the northeast where a significant plurality of customers using east regions probably live.