Elastigroup Managed Container Service strips away any need to provision, manage, or scale your infrastructure when running containerized workloads.
Container orchestrators (e.g.
Kubernetes) manage placement of your containerized applications onto instances, freeing you of the need to handle container scheduling and placement yourself. However, you are still tasked with provisioning, managing and scaling the underlying cluster on which your containers run. Spotinst MCS provides a management layer on top of common container orchestrators (Kubernetes, Amazon EC2 Container Service, Docker Swarm, and Nomad), which manages your cluster and servers for you.
How it works
With MCS, all you need to do is pick your container orchestrator, create your container definitions, and configure their system requirements (CPU and Memory). Spotinst MCS takes over from there, keeping your application highly available while exercising smart container packing to optimize for costs.
With Spotinst MCS, you can run containerized workloads on the public cloud provider of your choice, such as Amazon Web Services, Azure, Google Cloud Platform, and Alibaba Cloud.
How does Elastigroup MCS help me reduce my compute costs by up to 80%?
When you create a new Managed Cluster (
mCluster), Spotinst MCS automatically generates a Spotinst Elastigroup under the covers to manage the cluster and its servers for you.
Spotinst Elastigroup is an intelligent cluster management system that lets you save up to 80% on your compute costs while maintaining 100% availability, by maintaining a cluster of spot instances as the underlying infrastructure for your containerized applications. Elastigroup makes sure your cluster remains highly available at all times by leveraging predictive algorithms and smart resource allocation.
Does Spotinst MCS really manage and scale the underlying cluster for me?
Yes. With Spotinst MCS, you will never need to provision or manage a server or scale your cluster. Spotinst MCS leverages smart autoscalers to optimize your cluster size at all times.
The Spotinst autoscaler continuously monitor CPU, Memory and port availability, scaling your cluster up to accommodate for incoming tasks, and scaling it down when instances are underutilized. Finally, the autoscalers let you define and customize “headroom” for your cluster, making sure that your application always has enough capacity provisioned to handle spikes in traffic.
No need to manage your clusters or serversWith Spotinst MCS, you can focus on containers and forget about the underlying compute infrastructure. Spotinst MCS frees you from picking instance types, administering servers, and optimizing cluster utilization.
Auto ScalingSpotinst MCS uses predictive algorithms to Auto Scale your clusters based on your containers usage.
Native support for common container orchestratorsContinue to work with your existing orchestrator of your choice. Spotinst MCS natively supports
Docker Swarm, and
Securely run on your infrastructureThe entire cluster and servers are living on your own infrastructure. Get the best out of both worlds; no need to administer servers, and still run within your own VPC.
Monitoring and VisibilityGet a beautiful dashboard and instance & container level monitoring.
Save 80% on your compute costsSeamlessly run on Spot Instances and save up to 80% on your compute costs while maintaining 100% availability using Predictive ML.
Using the Spotinst MCS we achieved over 20% of cost improvements. A big win for Vero!Tu Hoang, Vero Vero
We have been using Spotinst in the production environment and managing the Elastigroups is a pure pleasure.Pawel Kobylak, Sr. DevOps Engineer UXPin
Spotinst MCS was the nirvana we were looking for in managing our complex EC2 Kubernetes environmentsShane Savoie, Chief Architect Ticketmaster
MCS for Amazon ECS
ECS Fargate experience on Spot instances, and without losing visibility into the underlying instances. MCS listens to ECS task events that report Port, CPU, and Memory utilization, and leverages this information to ensure optimal task packing on your
MCS natively supports all ECS task placement constraints to provide you with maximum control over your task placement strategy. With MCS, you never need to think about which AMI to use, or how to update the ECS container agent on your instances, as MCS will automatically install the latest ECS-optimized AMI, and make sure the version you are running is always up to date.
MCS for Kubernetes
Run Kubernetes on Spot instances without having to manage the underlying servers. MCS listens to Kubernetes scheduler events that report CPU and Memory utilization, and leverages this information to ensure optimal pod packing on your
mCluster. MCS natively supports all Kubernetes pod placement constraints such as labels, affinity, and anti-affinity to provide you with maximum control over your pod placement strategy.
MCS for Nomad
MCS for Docker Swarm