The Cloudcast #185 - Masters, Minions and Pods - Kubernetes 101

• Kismatic Website - http://kismatic.io/
• Kismatic on GitHub - https://github.com/kismatic
• Kubernetes on GitHub - https://github.com/GoogleCloudPlatform/kubernetes
• Intro to Kubernetes (tutorial) - https://www.digitalocean.com/community/tutorials/an-introduction-to-kubernetes
• Kubernetes homepage (Google) - http://kubernetes.io/

Topic 1 - Let’s talk a little bit about your background and why we asked you to come discuss Kubernetes tonight.

Topic 2 - We’re all familiar with Docker at this point, and generally familiar with the underlying container technologies. So where does Kubernetes fit in? (who runs it? what’s the input to the scheduler? what does it use to track resources at the host level? does it assume all machines are the same?)

Topic 2a - What makes Kubernetes easy to use and hard to use?

Topic 2b - Does it use/assume all the native container management tools, or does Kubernetes do some of that tool?

Topic 3 - Let’s walk through the basic concepts and suggested best practices around things like #apps/container, tagging and pods.

Topic 4 - Since Kubernetes came from Google, every just assumes it deals with scale well. But how does the scaling of that control plane work? Is it a single data-center view, multi-data center or smaller segments within a data-center?

Topic 5 - What Google-specific assumptions are built into Kubernetes that might not be broadly applicable to other companies?

Topic 6 - What are some of the common applications that companies use to get started with Kubernetes?

The Cloudcast #184 - Streaming Analytics for Distributed Applications

Projects of the Week - None this week

• SignalFx Website - https://signalfx.com/
• SignalFx REST API - https://support.signalfx.com/hc/en-us/articles/201270489
• TheNewStack covers SignalFx launch - http://thenewstack.io/signalfx-a-saas-to-monitor-apps-at-any-scale/
• Ben’s Blog - http://www.bhorowitz.com/the_past_and_future_of_systems_management

Topic 1 - Came out of Stealth recently (3/12), Give a quick overview of the company and the problem you are trying to solve… Given what SignalFx offers, it’s important to understand the people behind it. Let’s start with the background of the team - lots of large, webscale, distributed system background. [how much is “productizing lessons learned”?; how much is “the will be different in 5yrs”?]

Topic 2 - What does streaming analytics mean? Why do companies care about getting analytics faster? Why build an analytics engine to solve a monitoring problem?

Topic 3 - You mention (intro video) that you’re a company that builds services for distributed systems, which are run by product teams, not IT. You were previously at VMware. Can you talk about the different mindset those product teams have vs. IT teams, especially how SignalFx takes their ideas and feedback?

Topic 4 - Walk us through how your customers interact with your service? Where do metrics come from (app, message queue, etc)? How do you secure that API interaction? How are metrics different from logs or events?

Topic 5 - In the same vein as the shift from IT to the Product Groups, your co-founder mentions that Developers are closer to production than ever. What does that mean to the evolution of tools and overall psyche of application developers?

Topic 6 - You mention that SignalFx “double purposes as a Application Intelligence solution”. We’ve been watching lots of interesting SaaS applications  emerge that tend to have a more singular purpose (Logging, PerfMon, AppIntelligence, etc.). Are you hearing from customers that some consolidation of functions is needed?

The Cloudcast #183 - Container-Centric Application Deployments

Topic 1 - It’s unusual for us to have guests from different companies, but your stories have commonality. But let’s talk about both of your backgrounds (and company backgrounds) first.

Topic 2 - When I was watching this video of Khash (Cloud 66) at this Hacker News meetup in London, it looked to me like a concept I call “unstructured PaaS”, which is sort of a DIY PaaS, with the best-of technologies.

Topic 3 - We’re curious to learn more about ContainerNet, that is the backbone for the container networking of Cloud66 (using Weave technology) and how it really works.

Topic 4 - Both of you are at the forefront of this transition of container-centric application deployments. Where do you see the maturity in the market and what are the next big opportunities?

Topic 5 - You both seem to believe in the model of modularity for these new architectures. Beyond “giving customers choice”, what are the big focus areas in building elements of these modular architectures?

Topic 6 - What are some of the tangible business advantages that you’ve heard from customers when it comes to choice and modularity in this container-centric application model?

The Cloudcast #182 - Moving DevOps Forward with CI/CD

The Cloudcast #181 - Investigating the ELK Stack

The Cloudcast #180 - Understanding CoreOS Distributed Architecture

New ByteSized DevOps Podcasts - Logging, Monitoring and Application State

A few weeks ago we introduced the ByteSized DevOps Podcast series. Initial feedback from the community was very strong, so we've decided to do some more. We plan to release a few every week or two. Let us know what topics you'd like to see covered.

Monitoring and Logging


Stateful vs. Stateless Apps