Hello hello hello! This is the last edition of the (very short) Hey, Sys experiment.
I wanted to blog about code, in a consistent fashion, for an extended amount of time. I managed to do that for 8 weeks! Now on to other things.
What things, you ask?
I have been accepted into an accelerator program to propel my side-business Madjé’s Cookies and get into local stores. Since the business will be what I focus on outside of work, it makes sense for me to blog about it, my learnings, and how I’m handling juggling it with a full time job.
I will post about my weekly discovery as I tackle flavor development, branding and packaging, social media, and so on. Stay tuned!
Until next time, in a blog post near you.
xoxo,
Pam
Welcome to this week’s edition of Hey, Sys, the most random systems engineering series, where the puns are bad just like the economy.
Each week, at least for now, expect the following two segments:
Let’s jump in!
With the upcoming Product Team’s annual hackathon, the systems team has been busy doing some maintenance work on scaffolds for Rails and Express applications. …
Hello, hello, hello, and welcome to this new episode of Hey, Sys. I’m your host Pam Yam and I’m here to take a break from this hot and damp girl summer.
It’s honestly been a bit of an emotionally tumultuous month but I’m grateful for constants like Planck’s, Ethiopian food, and books.
Let’s dive in.
This week was a potpourri of tasks:
To honor my less than optimal mental state, let’s keep things lite shall we? Let’s (try to) explain caching. …
After frolicking in the Pacific Northwest with my cousin for a week, I returned to work on this past sweltering July Friday. To my surprise and, obviously, joy, a new bug was waiting for me!
Okay, time to dive in!
Our team runs sandboxes on K8s. Currently we have two different ways to keep them secure and away from public eyes: our VPN and basic authentication.
Because the aforementioned bug is related to the basic auth, let’s look at how it’s been setup, then dive into what I think is causing the issue.
How it works:
We use nginx
as our ingress controller. As such, to enable basic authentication we’ve have to add the following annotations to our ingress…
Hello hello fellow human. I welcome thee to the shortest edition yet of the Hey, Sys weekly…. blogpost? Newsletter? Technical journal? Who knows?!
I’m writing this on the way to Mount Rainier! Some might argue that I should focus on the scenic drive instead of writing this. I can’t say that I disagree! Yet, here we are.
Let’s jump in.
The kubernetes services have finally been exposed! We did it by adding Load Balancers K8s services that spin up elbs on AWS.
There was one more problem: because their creation was automatic, the load balancers had very unfriendly, long, and random hostnames. We remedied that by adding CNAME records to point our friendlier custom addresses to the equivalent load balancers. …
Hi hello, time for another one of these.
The intense humidity that is plaguing Washington DC is melting my braincells so there’s no time for chatter and lollygagging. Let’s dive into this week’s topic: Load balancers.
I know this week was supposed to cover Skaffold, but I changed my mind. I’ll do that another time.
Last week, I briefly touched on what I’m working on: exposing the redis
, mysql
and elasticsearch
k8s services while respecting my team’s cluster setup.
I thought I had found the solution, namely adding an ingress per service but alas, nope.
A quick Google search revealed services of type NodePort
and LoadBalancer
were usually the way to go with k8s so I tried both. The NodePort
didn’t work, but the LoadBalancer
did, which was promising! …
Here we are! Another week means another edition of more strange, and grammatically incorrect, content that is…. part newsletter, part tech-journal? WHO KNOWS.
Let’s dive in.
I was originally going to discuss Skaffold, but nope. I have decided to read a bit on port-forwarding. Let’s take a step back and discuss why.
The Vox Product Audience team asked us to enable connections from their development environment to services running on Kubernetes. The services in question are elastic-search
, redis
, and mysql
.
Once I was past the initial burst of panic triggered by the task’s novelty and urgency, I decided to set up an external hostname for these services via an ingress. …
*Taps mic**
Is this thing working??
A day ago, I promised myself I would stop doing things that are only motivated by this weird need to perform that this social media world has created. Thus, I will not put this “out there”, I will post this incognito, and secretly hope that someone stumbles upon it and makes it go viral.
Don’t lie, you’ve had the same thought about that *groundbreaking* tweet you posted yesterday 👀 .
Oh wait, that’s right, no one else is reading this! I guess could technically ramble on about nothing? …
In the spirit of KubeCon and all the Kubernetes content I consumed, I thought I’d write a short explainer on Kubernetes. This post aims to give a high-level overview of why and how Kubernetes is helpful. My hope is that it’ll give you enough knowledge to dive deeper into the docs.
Without further ado, let’s dive into it.
Kubernetes, pronounced koo-ber-nay-tees, is the container HBIC. I didn’t understand what the big deal was until I joined the infrastructure team. If, like me back then, you wonder why so many people rave about it, this post is for you.
innoDB is the current Mysql default storage engine and as such, is responsible for all of our SQL queries, such as insert
, select
, drop
, and more.
innoDB is a major key… but note that it isn’t the only storage engine out there. Before version 5.5, Mysql was using Myisam, which can still be set as the default when configuring databases. That said, innoDB provides new useful features that include:
About