Logz.io will be on the ground at Re:Invent 2019 in Las Vegas. Besides visiting us at our booth (which you most definitely should), we are also taking the stage. It’ll be one of the can’t-miss parts of Re:Invent this year. Here is a list of the best sessions you should bookmark.
1. Designing an open-source observability stack for Kubernetes
Logz.io co-founder Asaf Yigal will take the stage to tutorial a completely open-source stack for observability on Kubernetes containers and clusters. It will mention and analyze the best options for open-source telemetry, monitoring, tracing and of course logging options.
There will be an emphasis on newer demands on DevOps, particularly the expectation DevOps engineers can work security into their container orchestration. Kubernetes deployments are adding their own complexity to an already unprecedentedly complex environment for data management and log shipping. The talk will provide tips on integrating all these variables to make them work together in the most optimal way.
Knowing that other sessions will cover similar themes, Logz.io is going for an accessible presentation. Kubernetes is still a difficult topic even for experienced DevOps engineers, so helping people wrap their heads around how to best orchestrate their containers for monitoring and observability.
“Observability” will also see the buzzword treatment in 2020 and beyond, so defining it succinctly (including best practices) will be a priority in this presentation and further content and learning from Logz.io in the near future.
2. Building machine-learning infrastructure on Amazon EKS with Kubeflow
A team from UK startup Babylon Health (@babylonhealth) will demo Kubeflow (@Kubeflow), a machine learning-focused toolkit for working in Kubernetes. Yaniv Donenfeld, Jean-Marie Ferdegue, and Jeremie Vallee – Babylon Health’s senior bizdev director, director of platform engineering and AI infrastructure lead respectively – will cover why there was a need for a tool specifically tailored to machine learning workloads on Kubernetes. The presentation will also customizations, resource management and training job monitoring.
ML represents a different kind of bear to wrestle with on Kubernetes (and workloads in general). There are more intensive – and creative – demands when it comes to AIOps. Babylon Health is a client of Logz.io, so we are intimately familiar with their use case and the value of these tools when it comes to medical and health-tech applications.
3. Future-proof your career: Java developer to machine-learning practitioner
AWS technical instructor and computer vision expert Kesha Williams (@KeshaWillz) will catalogue her process of going from Java developer to machine learning expert, from learning Python and Tensorflow to AWS tools like Rekognition and SageMaker. Williams is also the founder of Colors in STEM (@colorsinSTEM) and an editor for DevOps e-magazine InfoQ (@InfoQ).
This might be one of the gems of the entire conference, as learning new skills can sometimes be an underappreciated goal of a lot of people already in the tech industry, even within the DevOps department.
4. Leadership session: Five new features of Microsoft and .NET on AWS that you want to learn
AWS VP Sandy Carter (@sandy_carter), Capital One’s software engineering director Peter Torres, and Expedia’s senior director Matt Parzybok will present how best to use AWS workloads on Windows, a very hot topic in the DevOps community. They will cover use cases that make extensive use of serverless architecture, containers and machine learning.
Windows can sometimes get forgotten behind the growing list of Mac and Linux resources or tutorials for cloud computing generally and monitoring specifically. Being able to move across OS platforms, especially at a time where integration is more important than ever, makes sessions like this a must.
5. Hybrid Cloud Dos and Don’ts
Hybrid cloud architecture should be high on your list of priority sessions if you’re on the ground in Vegas. Nutanix’s (@Nutanix) senior product marketing manager Sahil Bansal and senior director for product marketing Gil Haberman will tackle best practices and what to stringently avoid when it comes to automation and on-premises environments.
Despite AWS’s clear dominance, Azure and Google Cloud Platform (GCP) are on the rise. Additionally, as the old adage goes, diversify your stocks. How to split up workloads, like anything else, will have a set of best practices even if they’re not intuitive.
6. Building hybrid cloud architectures with VMware and AWS
PennyMac’s executive VP for systems architecture and ops, Nick Akl, with AWS Solutions Architecture Manager Andy Reedy, will give a rundown of important points for operating hybrid cloud architecture. During this session, they will focus (obviously) on AWS, but in particular how it works with VMware. This should be insightful regarding other combinations for hybrid deployments. It will specifically cover VMware workloads on Amazon S3 and RDS, as well as Elastic load balancing.
7. Deep dive: Kubernetes auto scaling
AWS Solutions Architect Chance Lee promises a detailed analysis of using Kubernetes’ autoscaling features, including HPA and VPA (which deploy or reallocate new pods when necessary). But, the session plans to present a full live demo of the Kubernetes Cluster Autoscaler (which handles deployment of nodes) as deployed on AWS (and requests anyone who wants to attend to please remember their laptop).
For all the best practices of setting parameters on your pods and nodes, and conversely how simple autoscaling sounds, there are a number of things you want to make sure are in place for a number of scenarios. The nuance of that issue will be exposed a little here, something that advanced AWS or Kubernetes users would benefit from recognizing by attending this session.
8. Moving to event-driven architectures
AWS VP Tim Bray (@timbray) is a popular voice in the DevOps community. This session, which will last an hour, will cover making the transition to an events-driven microservices architecture much, much easier. Bray plans to cover basics like and what kind of agents, event consumers and channels to have. He will also discuss what engineers should actually consider an “event,” how to publish these events, and which are worth monitoring.
“Events-driven” will likely be a big buzzword in 2020, if it isn’t already. For anyone getting into event meshes or dealing with platform integrations, this will be a pretty insightful experience. Bray is the kind of voice that serves as a good place to start for better understanding the fundamentals while diving objectively deeper than most others on this emerging hot topic.
An Open Smorgasbord of Info
As this list makes clear, there are a few priority topics you need to have on your schedule at Re:Invent. Observability, Kubernetes, security, and hybrid cloud architecture are common motifs on this list. Re:Invent will have literally hundreds of sessions, but these hone in on the big topics when it comes to logging and application monitoring. See you in Vegas.