Forem: Ali

How to do Subdomain URL Redirects with Vercel Hosting

Ali — Tue, 25 Jan 2022 22:35:25 +0000

I remember the first time my friend Ash showed me Vercel (at the time it named Now / Zeit). I was blown away. Since then, I’ve been using it to host my personal blog and personal website - both ali.dev domains.

One of the things I’ve been meaning to do over the past year is create subdomains for quick branded forwarding to different websites. There isn’t a straightforward tutorial on how to do this, so here we go!

What you’ll need

Vercel Account
Github Account
Custom Domain

This tutorial assumes you’ve already purchased your custom domain, and are managing it via Vercel. If this is not the case, please check out this document on custom domains on Vercel.

In this project, I will show you how I created a redirect for the subdomain, live.ali.dev, to my Twitch channel - twitch.tv/endingwithali. I actually wrote this blogpost live on my stream, using an Oculus Rift! I've included a clip of it below.

To write this blog post, I refered to this Github issue that was opened on Vercel's repo by SeinopSys.

Create a basic GitHub repo for the domain
Initialize the project locally - it won’t need much in it.
Create a document in the local folder called vercel.json - this is the only file you will need for redirects.
In vercel.json, add the following code:
```
{
    "redirects": [
        { "source": "/", "destination": "[https://twitch.tv/endingwithali](https://twitch.tv/endingwithali)" }
    ]
}
```
This JSON object says that every-time the ‘/’ domain is visited, Vercel will redirect the visitor to the destination URL. I wanted all visitors of the page to be immediately redirected to my twitch channel hence, "destination":"https://twitch.tv/endingwithali".
Commit code to the Github
Run the vercel command in your terminal to create a new vercel project. Make sure it’s a new project, not linked to any pre-existing project
Run vercel --prod to push your website live.

At this point, your redirect will be live, but not have the proper domain. In the next steps, we will set up the domain.
Go to your Vercel Dashboard ui (online, click into the project)
Connect your Github repo to your project by clicking:

We are connecting the project to the Github repo, because Vercel has it such that every time a commit is made to your repo, it’ll update the website. This makes it easier to update the URL as needed.
Next, click into settings > domains
Add the domain you want to redirect - for the subdomain, you can just type in the subdomain. The main domain should be maintained / controlled by the Vercel name servers. Click here to learn how to set that up.
Congrats you now have a domain redirect!

Go forth and redirect all the pages!

Bonus Video

This was originally posted on my personal blog, blog.ali.dev. If you want to watch the bonus YouTube video, you'll need to check it out over there ;)

Public Speaking Class with Josh Constine

Ali — Thu, 20 Jan 2022 18:30:00 +0000

I recently had the chance to sit down in a small setting with Josh Constine, to learn about how to be a better public speaker and better moderator. I wrote notes voraciously, and wanted to share them with you:

you sound slower than you think when you’re speaking, but still try to speak a bit slower
- the slower you speak, the less filler words you use
- you sound more dignified
- your brain can keep up with the words you’re saying
use dynamic tone
- emphasize words that matter
- take pauses at end of sentences
“finish your sentence” - make sure you come to the end of your sentence
ask questions with a finite ending
- keeps speaker on edge
- put the context at the start of the question
  
  “Web3 is growing big... how do you get speaking gigs?”
  
  instead of
  
  “how do you get speaking gigs? now that web3 is blowing up, there are more opportunities to find talks”
never answer for your panelist - if they don't answer right away, let them have the time / give them the space to think
- it's not on you as the moderator, to rescue the speaker
as a speaker, keep the answer hard and specific to the question
- do not dilute the speaker’s question
- formulating good answers take time
small gestures on stage are actually really distracting → try and stay motionless
- a little bit of gesture / posture change as a moderator is ok, but fidgeting is really noticeable
when you ask tough questions, be sure to ask in an empathetic tone
- helps break down walls
as a panelist, have 3 points you want to keep going back to
- prepare beforehand of what these three points will be - have a strategy
- if you don’t know the answer to a question asked, you can lead back to these 3 points
- 3 points is easy for an audience to remember
if you’re being interviewed - prepare the interviewer with points you want to talk about
- helps shape the direction of the conversation
if you need a prompt / interview helper, do not use your phone
- better to have an analog prompt (piece of paper, notebook, index card, etc) to help with answering question

💡 interviewer pro-tip - read the last few tweets of the person you’re interviewing! It will help you be relatable, also might be able to tie into the talk

listen to previous works / talks of the interviewee → you can tie it into your questioning, and allows for deeper conversations / questions
- It doesn’t have to be the whole thing, just a small snippet
pre-write your jokes, even if they’re not funny
- the audience will relax because bad jokes are still good jokes
prep calls are counter productive
- vibe doesn't carry over to the interview
- audience wasn’t there, so if you refer to something from the prep call, audience wont understand the context
- email threads >>>> phone calls
recommendation: have 2-3x more questions than needed
- you will have flexibility and no stress of filling the time
- have list of questions ordered by themes
  - allows for organic flow of conversation
don’t give exact questions to speakers beforehand because they’ll write canned / PR friendly answers
- instead give topics and subtopics you expect to touch on
you can get people to talk about “thorny” or hard topics by asking them three times
feel free to use “critics have said ____ how do you respond” or “people on twitter have said _____, what do you think?” to ask hard/edgy questions
always know your next questions → no dead time on stage as the interviewer
if you have a large number of speakers
- start lightening round of questions to get speakers talking / riff
  - this also hooks in the listeners / audience
- later you can ask to share a little bit more about their background and to answer a question at the same time
- personal and biographical questions help audience relate

“Give a quick 10 second intro and then answer what’s your favorite DAO right now?”

get to the meat and potatoes of the talk in the beginning of the talk to be sure to give enough time for deep conversations and answers
don’t be afraid to shut down long winded speakers
- this is your duty as a moderator to the audience
- redirect to another person if they’re talking too much or back to a topic if they’re off topic
end panels with something future facing - “what's next?”
do recaps - Josh likes to do recaps at the ends of talks
- takes notes mid talk quickly (using analog means) to do recap for later
- helps keep audience’s memory fresh of what was covered
- very impressive skill to speaker bookers
- moderator or speaker, you should do this
  - josh literally will do this as a speaker by grabbing the mic and announcing things to recap
a good question ordering as a moderator
1. speed round
2. personal/autobiographical question
3. tactical question
4. challenging question if needed
5. future looking question
6. recap
twitter spaces - talks do better with closed speaker loops and then allowing external listeners to ask questions
breaking open shy speakers by giving them an easy question to start with

Q and A:

it took Josh 4 years to start getting paid
- how bad do they need you?
- as a reminder, you have a lot more leverage last minute for getting paid
- typically you want them to throw out the first number
looking for gigs → you can probably get paid if a conference has paid tickets and sponsorships
find your core piece of credibility - “i made this and that’s why you want me to speak”
- assemble your references / sources and have them easily available for people to find
the hardest speakers to find right now are keynote speakers
- find a topic you care about and turn it into a 20 minute presentation
- conferences want keynote speakers
when do you feel like you have mastery / competency to talk publicly about a subject?

💪 “Fake it till you make it”
how to get competency about a subject to publicly speak about it, quickly
- get an advisor network to give advice, future insights, and to ask what are the current heated topics / debates in the field
  - get experts to tell you their insights
  - read a few blog posts
- formulate a new opinion after absorbing all this information
good keynotes have stats, lenses, and insights about a topic

It was a really engaging discussion, and I'm excited to share these notes with you. Thank you Josh for taking time out of your busy schedule to talk to us!

Bonus Video

This was originally posted on my personal blog, blog.ali.dev. If you want to watch the bonus YouTube video, you'll need to check it out over there ;)

Explaining Kubernetes for Developers Who Know How to Code

Ali — Fri, 14 Jan 2022 00:22:50 +0000

You know how to program - you're in your first or second year of your first software engineering job. You've mastered the basics and are excelling at your job, but something's been bothering you. You keep hearing about this thing called Kubernetes being thrown around by another team. Kuber...netes? What does it mean?

After the third or fourth time you hear it, you decide to look it up. You get what it does, but you don't really get what it does.

Don't worry—I’m here to help.

Before we get started, I highly recommend reading this article by fellow Relic John Withers - it's an excellent primer of basic colloquialisms and technical terms used when diving into the world of Kubernetes.

Hi, welcome to the container store.

Most starter tutorials will explain how to deploy a basic "Hello World" server in the language of your choice. But it doesn't really do much except return an endpoint that says "Hello World.” What do you learn from that? Why is it always a web server?

Let's take a step back and remember that any type of application can be containerized and deployed using Kubernetes. Even the hit game from 2002, Petz: Dogz 5, can be run using k8s.

Does it actually make sense to put Petz: Dogz 5 into a container and deploy with Kubernetes? Maybe, maybe not. But say we wanted to run 10,000 instances of Petz: Dogz 5 or we wanted to access the game 10,000 times at once, all simultaneously. Then yes, using Kubernetes makes sense.

But Petz: Dogz 5 doesn’t use the internet. Kubernetes must require the internet, right? Not exactly. Connectivity is a must, but things like intranet allow for connection without using actual internet, creating isolated instances. Keeping it straight forward, while you can run things on k8s that do not need internet access, you’ll need connectivity in order to access the k8s cluster.

Memenetes

Kubernetes has been the subject of memes...a lot. Why? Because in reality, setting up Kubernetes is hard. It's confusing and it takes a long time to get right.

However, once set up and configured properly, it's a powerful tool that can save developer time and application uptimes.

So when should one start using Kubernetes? And why?

Who? What? Where? When? Why? of Kubernetes

Who? + What?

Technically, anyone can run anything using Kubernetes. Usually, you'll find it being utilized by DevOps teams to help scale their company's services to meet incoming demand.

Where?

Kubernetes is run on a server (managing it yourself is known as “rolling your own Kubernetes”) or handled by vendor or cloud service provider as a managed Kubernetes.

When?

Technically, you can start using Kubernetes on day one of building a project, but it doesn't matter when you start using it. The choice is up to the developers and the DevOps team.

Why?

The ‘why’ question is where it gets a little more interesting.

Without Kubernetes, launching an application of any kind requires three different parts:

A compiled version of the code: provided by you and built in a form that can be easily run (such as binary).
An operating system: this might be your computer or something you set up manually, but could also be a container or use a buildpack.
A start command: The command used to execute your codes such as a binary or run statement.

Kubernetes x Docker: Best Friends Forever

Whenever you hear Kubernetes, you're probably going to hear Docker in the same sentence. Kubernetes and Docker are like peanut butter and jelly—they're a perfect pair. Docker combines the three parts required to launch an application into one command. This makes it easier to deploy (or start) new applications through the creation of containers.

Containers are a critical part of the Kubernetes metaverse because they're the smallest building block of Kubernetes orchestration. Just like how atoms are the smallest bits of matter in the universe, pods are encapsulations of containers. Pods are considered a unit of one or more containers. All the containers in a pod will be co-located on the same operating system.

Scale

Kubernetes can help make scaling your application easier. It's a distributed system that can automatically spin up new pods to help balance the load of incoming requests.

For example, you've just shared your side project on ProductHunt and HackerNews. As the creator, you're responsible for maintaining your side project's uptime and making sure that it works. It's starting to pick up popularity, and you see that your previous configuration is starting to max out at its ability to stay running. Did I mention that your project just blew up on Tiktok, and you've gone from 100 visits per day to over 10,000? It also just hit top of HackerNews - everyone is talking about it.

With the influx of users trying out your side project, you need to scale - fast.

Prior to implementing Kubernetes, you would need to go to your hosting site (either via a website for the cloud or on premises for non-cloud servers), manually provision the servers, run the containers, and set up the reverse proxies to handle incoming router traffic. With Kubernetes, this process is mostly automated. Since you were conveniently creating this side project to try to learn Kubernetes, you were prepared. To scale with the huge influx of demand, you're able to quickly update your Kubernetes config file to begin rolling out scaling changes for your side project.

By the way, New Relic 💙's Kubernetes

If you've just set up your first Kubernetes cluster or are looking to understand the Kubernetes cluster you've been running, look no further. I'm a visual learner, and having a tool like New Relic is critical for helping create awesome visualizations for your clusters. Learn more about our visualization tools and sign up for a free account today.

We've done it

As a developer, it's easy to forget to slow down, to get frustrated when learning new topics, and to gloss over asking why when one is confused. Hopefully, by reading this article, you've gained a bit more practical knowledge about Kubernetes and why you keep hearing about it everywhere you go.

Next Steps

Learn more:

Here are some excellent resources I'd recommend checking out if you're keen on learning more about Kubernetes from a highly technical angle:

FAANG is Over. It's a FANMAG world now. (The History of FAANG)

Ali — Thu, 10 Jun 2021 14:59:09 +0000

I was watching Cassidy Williams' Twitch stream recently, where she was talking about FAANG companies:

These clips inspired me to make this meme, which as of writing this post, is still getting lots of engagement across the Twittersphere:

// Detect dark theme var iframe = document.getElementById('tweet-1400561909754105867-654'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=1400561909754105867&theme=dark" }

As I started to see engagement roll in, as well as sharing it across some meme channels, discussions began to breakout about FAANG, the origins of FAANG, and some people not even knowing what FAANG is.

I also became curious: Why is Microsoft not included in this ubiquitous slang?

G

Before giving a full definition of FAANG, a detour.

Here's the google results when I search "FAANG":

Wait a minute.... let's enhance this....

(Did you know this is called a Google Knowledge Panel?)

uhh... are you seeing what I see?

⚡️ENHANCE⚡️

Uh.... Awkward....

What is FAANG?

FAANG is an acronym standing for Facebook, Amazon, Apple, Netflix, and Google (Amazon and Apple are interchangeable in order).

So that picture above, says Microsoft and not Netflix....

According to Investopedia, the concept of FAANG can be attributed to Jim Cramer of CNBC's Mad Money show:

The term was coined by Jim Cramer, the television host of CNBC's Mad Money, in 2013, who praised these companies for being “totally dominant in their markets”.

If you want to read more about Cramer's prediction of FANG stocks in 2013, I found an article expanding more about his / his team's thought process.

The term has been coopted by the tech community to describe its illustrious companies - the largest, most renowned companies that represent the industry across the globe in every trade.

Why isn't Microsoft part of FAANG?

Big Tech... is BIG. If you ask someone in the Bay Area what are the big tech companies they'll probably say "Google, Amazon, Apple, Microsoft, and Facebook".... Yet, the term FAANG is still used to quickly refer to the biggest tech companies in the world.

Asking the original question: Why wasn't Microsoft originally included in the acronym?

Mad Money is a talk show on CNBC dedicated to help to give stock advice to the average person - someone who doesn't have the wealth to invest in hedge funds, access to financial advisors, etc. As a show, Mad Money aims to democratize stock investing. In such context, Cramer brought up the concept of FANG as a series of stocks to explore if you're interested in investing in the technology sector.

At the time of the recommendation, Microsoft was still under the leadership of the infamous Steve Ballmer - the successor to Bill Gates, and early employee of the Microsoft. Microsoft did a great job growing with Ballmer as chief dog at Microsoft, but they definitely missed trends with outstanding strike outs.

Things weren't looking good at Microsoft, from a trend perspective, and stock experts at Mad Money probably were not keen on its continued success and excluded it.

Chart showing Microsoft, Google, Facebook, Netflix, Apple, and Amazon stock prices over time between April 2012 and June 2014
Source

Can FAANG Change?

If we look back to Investopedia we learn something important

Originally, the term FANG was used, with Apple—the second “A” in the acronym—added in 2017.

Apple wasn't even added until 2017.

Facebook, Amazon, Netflix, Google became the FAANG we know today.

Source

Over time, FAANG grew to no longer describe the noted companies in stock, but the noted companies in reputation and success. It has become so colloquial - so much so that I've noticed that many in people who work tech don't know the origin of FAANG - only that it represents major tech companies.

The term has been coopted by the tech community to describe the illustrious companies in the tech industry - the largest, most omnipotent companies that represent the tech industry across the globe in every industry.

So why can't we make a term that better portrays a more accurate collective opinion of the tech industry?

At this point, using FAANG to describe just the official names of the abbreviated companies isn't even right - in 2015 the larger Google company renamed itself to Alphabet.

FAANG and FANMAG

FAANG is a stock term - it's not reflective of the reputation, size, etc. of the modern tech industry. That's why we should use a term like FANMAG .

F acebook

A mazon

N etflix

M icrosoft

A pple

G oogle

Also, in case you were curious about the FAANG stock concept now, Cramer has come out in support of changing the abbreviation - remove Netflix, add Microsoft.

What do you think of FANMAG? Will you use the FANMAG acronym?

INB4:

PS - Netflix?

I'll say it - reputation wise, Netflix flys quite under the radar in the tech industry compared to the other companies of the FANMAG abbreviation. So, why should it be included?

What Netflix lacks in company clout, it makes up for in global usage. In 2018, Sandvine estimated that Netflix accounts for 15% of global internet bandwidth and the top application of global internet traffic. Its prevalence in the modern consumer society is undeniable, and a valid argument for defending its continued existence in the FANMAG abbreviation.

(header adapted from AlphaSense)

PPS - I stream on Twitch. Come be my friend ♡

How do web servers work?

Ali — Wed, 02 Jun 2021 16:35:08 +0000

Web Servers

They're mystical things... You just, initialize one for your project, and it handles the rest. But what is actually going on under the hood?

I got curious, so I asked my friend Jonathan Kingsley if he knew about 'em. Turns out he's the kind of guy who read the HTTP paper for fun, so we took to my stream and worked through building our own web server in Go.

Did I mention that the server is Mickey Mouse themed?

HTTP

HTTP stands for "Hypertext Transfer Protocol" - it was created as a part of the World Wide Web project in the early 90s.

Tim Berners-Lee, a British scientist, invented the World Wide Web (WWW) in 1989, while working at CERN. The web was originally conceived and developed to meet the demand for automated information-sharing between scientists in universities and institutes around the world. - CERN

This project has become the basis of how the internet works - it outlines the expectations of how to communicate data and information between servers. Your computer knows how to interpret websites because of the work of the World Wide Web project

What did we make?

We decided to make two iterations of a "from scratch" web server - the first one, outlined in this blog post, handles get requests to endpoints that are hard coded. In another blog post, I'll walk through how we built a web server that can handle more dynamically implemented endpoints.

How does the server work?

Main

Listening

At the basic level, a web server listens:

port, err := net.Listen("tcp", ":1928")

We create a Listener object, port, that listens to a specific traffic port/communication endpoint for incoming requests of a specific network protocol. This line of code basically starts the server - you can now receive incoming requests. In this example, we are expecting TCP style requests on port 1928. (Mickey Mouse was created in 1928, therefore we chose to listen to port 1928!)

Great! We're done! Right?

Not exactly....

We can now receive requests, but how do we handle and read them, and how do we send back responses?

Accepting

conn, err := port.Accept()

One of the generic functions that port has implemented is the Accept() function. Accept stalls/blocks until a new incoming request is seen by port - the incoming connection is returned as a Conn object. While the server is running, this function should be continually occurring - meaning that in code it is placed in an infinite for loop.

Yeah, you read that right - a time where an infinite loop is encouraged!

Handling

go handleConnection(conn)

We create a new custom function, called handleConnection - this function is where we actually start going through, interpreting, and reading the data from the incoming connection. Each time there is an incoming connection from another client (aka any external connection since we are the server), we accept it and spin off a new handleConnection to deal with that client.

We use the keyword go to spin off goroutines, which enables multithreading asynchronisity for the server. If we didn't use goroutines to enable concurrency, incoming requests that block will block everything on that thread, so you won't be able to serve as many incoming requests.

Code for main function

func main() {
    fmt.Println("Goofy: hyuck - booting up!")

    port, err := net.Listen("tcp", ":1928")

    if err != nil { //failed to set up server
        fmt.Println("Mickey: Oh no Goofy! It looks like there was an error starting up the server! ")
        fmt.Println(err.Error())
        return
    }

    for {
        conn, err := port.Accept() 

        if err != nil { //client failed to connect with server
            fmt.Println(err.Error())
            return
        }

        go handleConnection(conn)
        fmt.Println("Welcome to the Mickey Mouse Web House!")
    }
}

Printing Twice

fmt.Println("Welcome to the Mickey Mouse Web House!")

When you run the code above, this print may occur twice. When browsers execute a request from a server, they execute separate requests for both the favicon of the page, and the contents of the page.

handleConnection()

Anatomy of an HTTP Request Message

According to the specifications of HTTP, there's required information that must be passed in a specific order in order to be properly digested by the receiving side.

GET / HTTP/1.1\r\n
Content-Type: text/plain; charset=UTF-8\r\n
Content-Length: <length>\r\n
\r\n
Hello World!\n

Note: each one of these lines end with \r\n - when you're reading the incoming request, this allows the buffer to know where each line of the protocol ends. Basically, it helps separate new lines.

GET / HTTP/1.1\r\n

You can take this at face value. It tells the server the type of request incoming, the request target (the endpoint trying to be reached), and the HTTP version.

If you want to read more about the anatomy of HTTP Messages, check out this blog post by the team at Mozilla.

👋 I want to point out Content-Type header - up until I started working on this project, it didn't hit home as to why setting and checking Content-Type in HTTP requests was important. But now, I understand that each HTTP header acts almost like a basic if-else check point. Without setting the Content-Type specifically, the interpreting server won't know what to do or how to process the contents of the request. Processing each kind of request is basically hard coded.

Basically, always check your Content-Type header if you're sending a request, and if you're processing/receiving a request, be sure to specify the expected Content-Type in your documentation to reduce frustration on both ends!

Anatomy of an HTTP Response Message

This is what an HTTP response message is expected to look like:

HTTP/1.1 200 OK\r\n
Content-Type: text/plain; charset=UTF-8\r\n
Content-Length: <length>\r\n
\r\n
Hello World!\n

Let's break it down.

HTTP/1.1 200 OK\r\n

This line tells us what version of HTTP we are using, the status code, and the text associated with the status code

Content-Type: text/plain; charset=UTF-8\r\n

This header tells the receiving client the type of incoming content so it knows how to handle it.

Content-Length: <length> \r\n
\r\n

The value for content length is important, because it helps the receiving side know when the receiving message has been delivered in its entirety. It ends with double \r\n - the second one is a blank line used by request processors to note the start of the message body.

Hello World!\n

This is the message body!

Reading incoming requests

request, err := bufio.NewReader(connection).ReadString('\n')

We read the information from the incoming requests using a bufio object based on the incoming connection. To keep the server as simple as possible, we're not even going to check what kind of request is coming in, only what endpoint is being requested. This allows us to only have to read the first line of the incoming HTTP request - we don't care what the rest of the request says.

requestParts := strings.Split(request, " ")

We split the incoming request into parts and check...

if requestParts[1] == "/clubhouse" {

which endpoint is being requested. For this server, we're only accepting one endpoint: clubhouse.

message := "if goofy has a dog, and goofy is a dog....????"

connection.Write([]byte("HTTP/1.1 200 OK\r\n"))
connection.Write([]byte("Content-Type: text/plain; charset=UTF-8\r\n"))
connection.Write([]byte("Content-Length: " + strconv.Itoa(len(message)) + "\r\n\r\n"))
connection.Write([]byte(message + "\n"))
return

To send a response, we use the Write function of the connection object - but before we do that, we need to send our required headers (as outlined above and in the hypertext transfer protocol).

If the incoming request is not looking for /clubhouse, then we send back a 404 error.

connection.Write([]byte("HTTP/1.1 404 Not Found\r\n"))
connection.Write([]byte("Content-Type: text/plain; charset=UTF-8\r\n"))
connection.Write([]byte("Content-Length: 0\r\n\r\n"))

The final thing that happens before we finish the method is that we must close the connection

defer connection.Close()

In reality, this was the first thing we did in our handleConnection function - If you're new to Go, defer is a keyword utilized to describe a function that should be executed when the main function completes, no matter where it ends.

Code for handleConnection()

func handleConnection(connection net.Conn) {
    defer connection.Close()
    request, err := bufio.NewReader(connection).ReadString('\n')

    if err != nil {
        fmt.Println(err.Error())
        return
    }

    requestParts := strings.Split(request, " ")

    if requestParts[1] == "/clubhouse" {
        message := "If Goofy has a dog, and Goofy is a dog....????"

        connection.Write([]byte("HTTP/1.1 200 OK\r\n"))
        connection.Write([]byte("Content-Type: text/plain; charset=UTF-8\r\n"))
        connection.Write([]byte("Content-Length: " + strconv.Itoa(len(message)) + "\r\n\r\n"))
        connection.Write([]byte(message + "\n"))
        return
    }

    connection.Write([]byte("HTTP/1.1 404 Not Found\r\n"))
    connection.Write([]byte("Content-Type: text/plain; charset=UTF-8\r\n"))
    connection.Write([]byte("Content-Length: 0\r\n\r\n"))
}

Final code on Github.

Serving Endpoints

I was surprised to learn that servers essentially just reading and parsing string. After years of working with them, I genuinely thought servers were way more complicated and that a simple example like this would be much more complex.

Right now, our server can only handle one endpoint /clubhouse - which has a hardcoded response. What about dynamically implemented endpoints? Ones that are not defined directly in the server code - don't worry we've got that too. Blog post coming soon.

BTW, if you're curious if New Relic works with servers... yes it does. Click here to learn more.

Special thanks to Jonathan Kingsley for working on this project with me! Having a mentor to help walk me through this process and has made it very digestible! I am very grateful!

Wish you could have seen this learning live? I stream coding and other fun tech things on Twitch almost every day. Come hang out! See y'all soon!

Mickey Mouse

Yeah... while this doesn't use Mickey Mouse for any kind of analogy, I've been working on my Mickey Mouse impression and the entire time Jonathan and I streamed this project, I kept doing my impression, hence the use of Mickey Mouse.