7 ms·

> On the darker side, inefficient scaling (due to uneven daily/yearly traffic distribution), overly complicated deployments (delivering code to many boxes quick

> On the darker side, inefficient scaling (due to uneven daily/yearly traffic distribution), overly complicated deployments (delivering code to many boxes quickly is hard), and fragile service discovery (have you tried running consul or zookeeper at scale?) would lead to higher operational costs. Obviously, everyone's use-case is different but none of this has been my experience with VMs: - Both AWS and GCP allow you to define custom scaling policies for an instance group. - Deploying new code just means updating a template with your new image and pushing that out. Both GCP and AWS would then cycle the existing machines automatically to use this new image. - At least in our case, there is no need for Service Discovery; you just need to know where the Load Balancer is. We have moved some things to Kubernetes and now have to deal with inexplicable occasional failures. That and an everyday massive YAML headache. YMMV

awoimbee 4y ago
IMO it takes time to get up to speed but then it's pure bliss in terms of reliability. And don't go to yaml hell, use pulumi or equivalent.
- x86x87 4y ago
  > it's pure bliss in terms of reliability narrator: it is not. you're not trying very hard if you haven't hit a case where k8s will lead to you banging your heads on any available walls trying to understand a behavior in an edge case (here is one that made walls hate me: https://github.com/kubernetes/kubernetes/issues/96858 https://github.com/kubernetes/kubernetes/issues/96858 and I know of a half dozen developers that learned about this the hard way).
  - hanikesn 4y ago
    Looks to me like an example of aws not having good defaults put of the box and not an k8s specific issue. Having to specific a `sleep 30` shutdown hook on pods is quite annoying though and some built-in should be added instead.
    - shaklee3 4y ago
      exactly. if you read the bug it's clearly not a k8s issue. they cannot control aws jabbering elb bugs.
      - x86x87 4y ago
        You're missing the point. There is no cloud agnosticism and you need to understand the underlying cloud primitives although allegedly k8s isolates you from this.
        shaklee3 4y ago
        I'm not missing the point. The OP was specifically talking about a k8s bug that wasn't a k8s bug at all. You don't blame Linux for a poorly written application. If the cloud providers choose to add a bunch of uncommon options that has no counterpart in k8s that's their choice. but it doesn't mean k8s is a magic bullet that makes it so you don't have to do the bare minimum to understand the cloud providers.

awoimbee 4y ago

IMO it takes time to get up to speed but then it's pure bliss in terms of reliability. And don't go to yaml hell, use pulumi or equivalent.

x86x87 4y ago
> it's pure bliss in terms of reliability narrator: it is not. you're not trying very hard if you haven't hit a case where k8s will lead to you banging your heads on any available walls trying to understand a behavior in an edge case (here is one that made walls hate me: https://github.com/kubernetes/kubernetes/issues/96858 https://github.com/kubernetes/kubernetes/issues/96858 and I know of a half dozen developers that learned about this the hard way).
- hanikesn 4y ago
  Looks to me like an example of aws not having good defaults put of the box and not an k8s specific issue. Having to specific a `sleep 30` shutdown hook on pods is quite annoying though and some built-in should be added instead.
  - shaklee3 4y ago
    exactly. if you read the bug it's clearly not a k8s issue. they cannot control aws jabbering elb bugs.
    - x86x87 4y ago
      You're missing the point. There is no cloud agnosticism and you need to understand the underlying cloud primitives although allegedly k8s isolates you from this.
      - shaklee3 4y ago
        I'm not missing the point. The OP was specifically talking about a k8s bug that wasn't a k8s bug at all. You don't blame Linux for a poorly written application. If the cloud providers choose to add a bunch of uncommon options that has no counterpart in k8s that's their choice. but it doesn't mean k8s is a magic bullet that makes it so you don't have to do the bare minimum to understand the cloud providers.