Three Takeaways, "Learn Kubernetes in One Hour". Video in English and Spanish

I must confess that I didn’t expect the first workshop of the year (“Learn Kubernetes in One Hour”) to draw the crowd it did (198 attendees to be exact). Kubernetes has been around for a few years (Paul Bakker’s 2016 lessons learned is a great case study), but Kubernetes is still a hot topic that 33% of our attendees indicated as “Just Starting,” while 24% of attendees indicated that it’s “Purely a learning exercise.”

Poll from the workshop shows that a lot of eng teams are “Just starting” with Kubernetes despite it being around for a few years.

Many thanks to Christian Meléndez (Cloud Architect at Equinix) and Dave McAllister (Community Guy at Scalyr) for making the workshop a success. And most importantly, a big thanks to everyone who attended and signed up.

I’ve included three key takeaways from the workshop, along with the full recording of the workshop and access to GitHub repos below.

Takeaway #1: Kubernetes is easy to do on the cloud (but don’t force its adoption)

As tempting as it is to Kubernetes all the things, you have to start with “Why?” Yes, it’s the revelation of the century, but in all seriousness, you have to figure out if Kubernetes fits what your team needs.

The good news is that a lot of teams have figured out deployment, logging and monitoring. The bad news is that a lot of teams have figured out deployment, logging and monitoring (that work specifically for their environment). There’s really no rush to shoehorn Kubernetes (unless it’s something you absolutely have to solve yesterday; if so, takeaway #2 helps).

Takeaway #2: Consider how to make things simple (Kubernetes is already complex)

The timeless adage “Keep It Simple, Stupid” applies to everything in life, including Kubernetes. The TL;DR of all of this is:

Abstractions make life “simpler”
Abstractions make digging harder

Takeaway #3: Monitor each layer explicitly

Yet another “No duh, dude,” but I cannot emphasize this takeaway enough (especially as it relates to takeaway #2 about simplifying as much as you can). One thing to keep in mind is that you have to collect logs and monitor at each layer. And, be sure you monitor your orchestrated environment and account for ephemeral containers. It’s easy to forget (real talk: sometimes at Scalyr we forget to account for ephemeral containers).

Workshop Recording

English version:

Spanish version:


Github Repos

Courtesy of Christian. Follow him on Twitter, and check out his writings.