Do You Really Need Kubernetes (K8s)?

ByErma F. Brown

Jun 30, 2022 , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,


Illustration showing the Docker and Kubernetes logos

Kubernetes has a status for remaining difficult. When it’s a superb piece of software package, occasionally the added benefits may not outweigh the extra complexity, specifically for smaller businesses running easy deployments. Is plain aged Docker superior sufficient?

What Kubernetes Is Wonderful At

Kubernetes (K8s) is utilized with Docker pretty usually in qualified environments, and this can make it seem to be like it is an upgraded or much more arranged version of Docker. While this is true in some regards, it is not the correct reason of K8s.

Kubernetes is a “container orchestration procedure,” and it’s crucial to keep in mind what that means. The Kubernetes Command Aircraft handles all the facts of “orchestrating” which containers will run on which nodes. It separates the hardware from the software and allows your machines just run as employee nodes, which is preferably all they really ought to be.

Kubernetes great when you have a significant, production deployment that desires to be equipped to scale, maintenance, and deal with by itself routinely. When you’re taking care of hundreds of containers throughout many servers, owning to offer with load balancing and going containers from server to server can get bothersome. K8s handles this for you with excess configuration options.

Graphic showing Kubernetes cluster architecture
Graphic exhibiting Kubernetes cluster architecture

Kubernetes is also related to infrastructure as code (IaC), yet another helpful concept. It allows you management your solutions working with configuration documents that can be tracked through Git. This can substantially boost your deployment processes by letting it to be tracked alongside branches, examined, versioned, and rolled again if required.

What Kubernetes Isn’t Great At

The principal downside of Kubernetes is that it is notoriously intricate. This may not be a substantial issue if you are now acquainted with it, but in any scenario, it provides excess time and complexity to having your programs up and functioning. It is a great tool to study, and a good ability to have. But it is effective superior with larger sized deployments, and if you are running only a handful of servers or a couple containers, the added effort of Kubernetes may not be value it.

Kubernetes is a middleman, but you can generally go old-university and choose the function of orchestrator you. The pleasant issue about Docker is that it will make the process of creating and shifting programs all over simple with containerization. Generating a new server is reasonably straightforward, and you could have a new EC2 occasion from AWS managing your container in a way of minutes.

Without Kubernetes, the only additional problem you have is needing to operate commands manually on your servers with the docker CLI or docker-compose API. If you have to have to add a new container, or transfer a container to a different node, you will have to have to run these commands again. This can be an issue if, for example, you experienced unexpectedly superior visitors and necessary to launch more situations mechanically. K8s would address that dilemma. But in lots of scenarios, your targeted visitors might be comparatively steady, or at the very least predicable on a week-to-7 days foundation, particularly for backend services that do not seriously have to have to vehicle-scale.

Effectively, Docker alone currently increases the system of running moveable purposes by a big total. It is a massive advancement above getting to put in purposes the previous way. Kubernetes improves upon Docker a little bit, and automates some tasks essential for large networks, but provides a large amount of complexity.

The worth of it will depend on how much time it truly will save you. If this complexity is acceptable to you and ends up conserving you time, you need to use K8s, but if it does not, you should not waste your time with it—especially when you can often basically swap to it afterwards on if need to have be.

What Need to You Use Alternatively?

It’s fantastic to just take a step again and appear at what you are genuinely looking at Kubernetes for. If you want your community to car-scale, manage your resources quickly, and start pods on unique servers based mostly on configuration, Kubernetes is a good software to master. If you’re concerned about the complexity, cloud companies like AWS do have managed Kubernetes products and services like EKS, which can deal with some of the dirty perform for you, primarily when it comes to administration and provisioning resources.

Having said that, if all you want is a way to control your containers much more conveniently, or car-update containers, there are considerably simpler resources out there for the task. They may possibly not be as entirely highlighted as Kubernetes, but they’ll be simpler and can conserve you time.

If you want to automobile-scale selected services, you are going to almost generally have to have to be chatting to a cloud provider’s API to provision new means. Kubernetes can deal with this for you for a lot of platforms, but providers like AWS, Azure, and GCP all have simple container expert services with vehicle-scaling options. AWS’s ECS company can easily be set up to auto-scale to meet up with superior demand.

Relevant: How to Get Began with Automobile-Scaling Container Deployments on AWS ECS

If you will need to auto-update containers anytime you push new versions to your Docker registry, you can use a company like Watchtower. It runs on a employee node as a Docker services by itself, with the Docker socket exposed, and watches for changes. As soon as a new container is pushed, Watchtower will restart the container. This will conserve you time operating restarts of containers, and is a good alternative to Kubernetes updates.

docker run --detach 
--identify watchtower 
--quantity /var/operate/docker.sock:/var/operate/docker.sock 
containrrr/watchtower

If you want to be additional structured in general with how you manage your containers, you need to consider location up a Docker administration GUI like Portainer. Portainer is a net-based GUI for taking care of Docker deployments. It is effective with various employee nodes, substantially like Kubernetes, but it lets you cope with the task of orchestration, location up containers on the servers you want.

Portainer is truly straightforward to established up considering that it’s packaged as a container itself. If you’d like to learn far more, you can read our guideline to setting it up and performing with it.

Relevant: How to Get Started out With Portainer, a World-wide-web UI for Docker





Source backlink