Experiences: Observing k8s control plane while transmuting DevOpsES

1. Dev+Ops: Culture Wars by Dustin McKenna Watts

Everyone has had that moment in their career at least once - it's 3pm Friday afternoon, everyone is about to head home, when you get a ping in Slack. Some guy from some feature team wants to push some changes out to production. Nice. After contemplating the nicest way to tell him his "mistake", you probably might even start to wonder why devs do that.

Spending the last 8 years as a developer, and recently moved into DevOps - I'm here to explain through my experience why Devs and Ops seem to be at odds from time to time. Why it can be challenging trying to break the cycle, and (for you devs out there), how some simple lifestyle changes may be the answer.

About the speaker: I first visited Serbia 3 years ago on a business trip with NCR, and it was then that I met my wife. I moved here shortly after meeting her and married in 2020. I am an advocate of open-source tech and spend a large amount of my free time researching and playing with tech on my at-home Linux-based lab. I am passionate, experienced, and motivated to innovate.

2. Experiences: Monitoring Production Kubernetes Clusters with Prometheus by Boris Petrovic

In this talk I will share my experiences with operating and monitoring on-premises Kubernetes cluster(s) with Prometheus. I'll shortly describe how Kubernetes control plane works and how to monitor it, also I'll talk about things I find useful in monitoring other Kubernetes workloads (pods, deployments, ingresses, etc...). I'll back this topic with real-world situations that United Cloud DevOps team has experienced so far (in production).

This talk can be useful for those who are responsible for keeping both managed and on-premises clusters up & running and for those who want to learn more about Kubernetes itself and how we can integrate Prometheus with Kubernetes to get self-service monitoring solution for other Kubernetes users.

About the speaker: I'm Boris, a DevOps engineer in United Cloud. I like to discover the dark corners of Linux, especially networking stack stuff. I hope I will figure out how that thing works at the end. I enjoy tasks that require low-level OS optimisations and, I like to know how and why something works as it works. Since monitoring can answer this question, I have spent a lot of time creating a perfect monitoring system. Later, I've realised that such a thing doesn't exist.
Proud owner of dozen projects that will never see the light of a day, because I prefer to start things rather than finish them. If you couldn't watch your favourite TV show on EON, there are great chances that I was guilty for that, because - "this line in Nginx configuration will not affect anything, enter".