4 ms·
I only have one tiny detail: Bursty traffic means that your cluster needs to be able to deal with the peaks. If you running an on prem Kubernetes cluster, then
by mrweasel 6y ago
I only have one tiny detail: Bursty traffic means that your cluster needs to be able to deal with the peaks. If you running an on prem Kubernetes cluster, then there is no savings, unless you can use the capacity for something else during non peak periods.
The scaling, and potential savings is a cloud feature, not a feature of Kubernetes.
- tstrimple 6y agoIt also enables you to achieve app density. Many of the companies I've been working with lately have large batch processes on nightly / weekly / monthly basis. For some reason, each job previously was setup on on-prem hardware dedicated per workload. Using Kubernetes and scheduling jobs to manage capacity has enabled us to reduce the number of on-prem servers substantially as part of the migration plan for the cloud.
- halbritt 6y agoPrecisely. In my experience hosting services in VMs on-prem I was able to achieve roughly 30% efficiency across 3k instances and hundreds of nodes. In my experience hosting microservices on k8s, I was able to achieve 80% efficiency across hundreds of nodes. Both were the result of a great deal of work to optimize efficiency. In this case, I use the word "efficiency" to refer to a blend of CPU and memory utilization.
- cat199 6y ago> no savings This only relates to hardware cost - tooling impact can provide cost improvements if it enhances productivity (not always clear for k8s case)
- aliswe 6y agoI would nitpick that and say "if you are running a cluster with a static number of worker nodes". If I would have a cloud hosted k8s cluster and don't configure any auto scaling, the situation would be equivalent to me buying x number of physical machines and hosting them as worker nodes on prem. No ability to scale to "bursty traffic".
- ec109685 6y agoIf the burst is for all apps at once, then yes. But if you need to handle temporary bursts of a single app among manh, then hosting them in a heterogenous cluster will give you benefit.