Thread by @tpschmidt_

📚 AWS 1x1 - 𝗘𝗖𝗦

Easily run, stop, and manage containers in the cloud.

All you need to know about tasks, task definitions, clusters & containers ↓

𝗧𝗵𝗿𝗲𝗮𝗱 𝗢𝘃𝗲𝗿𝗶𝗲𝘄 🧵

• About Docker
• Fundamentals
• Task Definitions
• Tasks
• Services
• Clusters
• Container Instances vs. Fargate

{ 1/17 }

𝗗𝗼𝗰𝗸𝗲𝗿

Before getting started with ECS, you need to understand Docker, because it's one of the basic building blocks.

Docker helps to create environments to run your application, regardless of the underlying operating system.

{ 2/17 }

This lightweight environment is called a 𝗗𝗼𝗰𝗸𝗲𝗿 𝗰𝗼𝗻𝘁𝗮𝗶𝗻𝗲𝗿 and contains everything that is needed to run your application, e.g. certain versions of a library or language.

You can even run multiple containers on the same machine.

{ 3/17 }

Containers can even communicate with each other when needed.

You may have already guessed: looking at a growing application, there will be soon challenges in managing all those containers: deployments, scheduling, scaling and more will be a tedious task.

{ 4/17 }

That's where you'll be in need of a 𝗰𝗼𝗻𝘁𝗮𝗶𝗻𝗲𝗿 𝗺𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 𝘀𝗲𝗿𝘃𝗶𝗰𝗲.

It's another abstracting layer helping you to easily manage your containerized applications and reduce your operational tasks.

That's where ECS comes in!

{ 5/17 }

Amazons 𝗘lastic 𝗖ontainer 𝗦ervice is a highly scalable & fast container management service.

It allows you to view & manage the state of your clusters from a centralized service. Easily schedule based on resource needs & availability requirements.

{ 6/17 }

Let's have a look at two different areas of responsibility:
• Lifecycle management of your tasks
• Actually running your containers

ECS does not actually execute or run your containers. It only provides the management pane for controlling your tasks.

{ 7/17 }

So finally to the important question: which services actually run our containers?

You can either pick from using 𝗘𝗖𝟮 𝗶𝗻𝘀𝘁𝗮𝗻𝗰𝗲𝘀 (that have the ECS Container Agent running) or 𝗙𝗮𝗿𝗴𝗮𝘁𝗲 Launch Type

So it's not ECS 𝗼𝗿 Fargate, but ECS 𝗮𝗻𝗱 Fargate.

{ 8/17 }

Let's dig into ECS' 𝗞𝗲𝘆 𝗧𝗲𝗿𝗺𝘀, which can be confusing at first but are crucial to understanding how it's internally working:

• Task Definition
• Task
• Service
• Cluster

{ 9/17 }

A 𝗧𝗮𝘀𝗸 𝗗𝗲𝗳𝗶𝗻𝗶𝘁𝗶𝗼𝗻 is a blueprint of your container. It includes things like

• the image to use (e.g. exakt ECR URL)
• CPU & memory allocation
• secrets & environment vars
• logging configuration
• exposed ports

{ 10/17 }

A 𝗧𝗮𝘀𝗸 is an actual instance that runs the containers that are defined in your definition.

You can run multiple tasks from the same definition if this is required (e.g. for redundancy or traffic demands).

{ 11/17 }

𝗦𝗲𝗿𝘃𝗶𝗰𝗲

As we can have several tasks for the same definition, we need some boundaries and management. This is where the service comes in.

It defines how auto-scaling & load distribution is handled and what are the minimum & maximum tasks.

{ 12/17 }

Auto-Scaling rules can be for example that a new task is started if the CPU load exceeds a certain threshold for a defined period of time - or when an existing task is gracefully stopped if the load is too low.

{ 13/17 }

A 𝗖𝗹𝘂𝘀𝘁𝗲𝗿 is a logical grouping of tasks or services which run on infrastructure that is registered to such a cluster.

If you're using EC2 launch types, it's a set of container instances you manage. If using Fargate, infrastructure capacity is provided by AWS.

{ 14/17 }

You can even provide your on-premise virtual machines as compute capacities for your cluster.

Additionally, it's possible to have a cluster with 𝗺𝗶𝘅𝗲𝗱 tasks hosted on Fargate, EC2 instances, or even external instances.

{ 15/17 }

ECS Container Instance vs. Fargate

If you have very high computation requirements, you should know that Fargate is way more restrictive regarding what capacities are possible within a single task.

• 𝘃𝗖𝗣𝗨𝘀: 448 vs. 4
• 𝗠𝗲𝗺𝗼𝗿𝘆: 26TB vs. 30 GB

{ 16/17 }

Even if you're a big serverless fan, knowing about ECS is crucial because you'll bump into it almost everywhere.

Considering the abstraction layer of ECS in combination with Fargate, it's considered a serverless technology.

{ 17/17 }

That's it for the fundamentals - thank you for reading!

If you've liked this thread, ♻️ or ♥️the initial post & follow me for more regular cloud content 🌤

Also, comment or drop me a message about topics you're interested in! 📨