2016 February

2016 Ignite Proposal: An ElasticSearch Cluster Named George Armstrong Custer


This is the epic tale of an elasticsearch cluster named “Custer”, named after the US Calvary Commander.

It didn’t start out with a name, it started out as our logstash cluster, but after experiencing how easy it was to search and find logs using logstash, our developers took it upon themselves to see what else it could be used for. This was our Battle of Little Big Horn.

In this ignite talk, attendees will laugh and learn as I explain:

  • the origin of our cluster
  • how it got it’s name
  • Where things went wrong
  • Steps we took (good and bad) to deal with daily crashes (a.k.a. Little Big Horn)
  • Lessons learned on using and scaling elasticsearch

Attendees will leave this short, 5 minute talk with the ability to identify the signs and symptoms of an over-utilized cluster. Specific thresholds that proved invaluable to us, such as JVM heap usage, memory utilization, shard movement, and index latency will be shared. They will also learn the tips and tactics on how to recover once a problem cluster is identified.


will_buttonWill Button is a DevOps Advocate for Trax Technologies. He works closely with Development and Operational teams to align resources and strategies, focusing on communication. When not at Trax, he can be found authoring lessons for Egghead.io and Pluralsight as well as speaking at tech events. In his free time he enjoys coaching at CultFit Four Peaks. I mean CrossFit Four Peaks.


2016 Speaker Proposal: Continuous Delivery in the Application Economy by Burt Klein

The shifting focus toward an application-centric world radically changes the traditional environment large enterprises have known for so long. The velocity of change has accelerated exponentially over the past five years or so, and this necessitates a different approach to delivering services.  CA will share common themes we see in Fortune 30 companies who have embraced Continuous Delivery along with their DevOps Transformation.

Burt KleinAs a Senior Customer Advisor/Technology Evangelist, within the Application Delivery business unit of CA Technologies, Burt is responsible for informing leading worldwide organizations, and alliance partners, on the value of Service Virtualization and DevOps within the software development lifecycle. Burt has been traveling the globe speaking to the revolutionary technology capabilities of Service Virtualization. He is also responsible for sharing the success stories of countless early adopters and actively participating with executive level steering committees that are focused on the large enterprises implementation of this new and exciting capability.

Burt joined CA Technology after retiring from Bank of America where he was a technology executive with extensive experience in all aspects of software development lifecycle. Most recently Burt was the executive responsible for Performance Engineering and Validation for the Global Consumer & Small Business Banking at Bank of America.

He possesses vast knowledge in the areas of performance and resiliency validation, application software development & integration, automation & productivity effectiveness, environments management & support, test data management, software configuration management and process definition.

2016 Speaker Proposal: I don’t belong here – Working through impostor syndrome By Jake Champlin

Being relatively new to the professional development space, I’ve struggled on a virtually daily basis with impostor syndrome. The feelings of “I’m not smart enough, I’m not good enough, I should just quit”, wash over me at various points in the day. Working with such young and brilliant people, who’s talents and skills seemingly tower over my perceived lack of intelligence, creates several instances where I am completely floored at the professional level of work that these amazing people churn out on a daily basis.

Jake champlin

While historically my initial reaction has been, “I should just quit, I’m never going to be that good, why even get out of bed?”, I’ve learned several ways to cope with these feelings whenever they arise. In this talk I will go over various ways that I have found to cope with impostor syndrome, and how I’ve learned to continually improve to meet my goals. I plan to go over the differences between reality, and what a person struggling with impostor syndrome experiences; as well as ways in which coworkers and people without impostor syndrome can help their fellow coworkers who may struggle with it.

Jake is a Site Reliability engineer at HashiCorp. Before working at HashiCorp, Jake was a Operations Engineer at Minted, and before that a Systems Administrator for a local datacenter. He is a co-maintainer on a few Open Source projects, and an Open Source advocate. Jake is a self-admitted emacs lover, Linux user, and has been known to hate using a mouse. When he is not maintaining uptime, Jake can be found riding his motorcycle with his beautiful wife, constantly fixing his Jeep, or searching for the best place to get pancakes. 


2016 Speaker Proposal: Fear and Loathing in Systems Administration


DevOps doesn’t matter.  I will talk about ways that Systems / Operations teams can have a better quality of life, while providing better service, and developing better relationships with their companies, all while invoking the “D” word a minimum of times.


“DevOps Doesn’t Work.”  Whether it’s from people who’s “DevOps Transformation” consisted of a renaming, or people who scoffed at the idea of letting development deploy to Prod more often, I hear this all the time.  And frankly, I’m tired of “What is DevOps” talks.

DevOps is a strawmanperson.  When a SysAdmin says “DevOps doesn’t work” or grumbles about giving access to Production to people, the real problem is that historically they have been outnumbered, overridden, and have not had authority aligned with responsibility.

In this talk, I will demonstrate why DevOps doesn’t matter, and that focusing on DevOps is missing the point.  I will offer practical solutions to:

  • the problems of historically poor relations with other teams
  • All of the responsiblity, none of the authority
  • Tragically slim Systems / Ops teams
  • Constant Firefighting

velocity_nyc_2014-headshotWaldo is a Geek, Problem Solver, Systems/Operations Engineer, and DevOps Enthusiast.  He tries to use his Viking Privilege superpowers for good, but is a flawed human being.  He aspires to be a better bio writer one day.

2016 Speaker Proposal: Misconceptions of Automation


Do you remember that time the janky script you banged out was discovered to be the lynchpin of Production? Have you ever faced an “Automate ALL-THE-THINGS” project? Has your management ever-so-helpfully suggested that you “just automate it”? In this talk, I’ll discuss the mental gap that exists between automation and other business software.


Automation, the panacea of tech. It is necessary to work effectively. It’s needed in order to ensure consistency, prevent mistakes, and free engineers up to do more interesting work.

Infrastructure and delivery code is – at it’s best – invisible. But when it fails, the janky thing that you whipped up to solve a problem so you could move on with life is discovered to be in the critical path of Production.

You should absolutely automate everything that you can, but there are a many assumptions that go into automation, particularly by those who aren’t doing the automation work themselves. More important than writing automation is setting expectations. In this talk, I’ll cover

  • No, automation won’t put your job at risk
  • Setting Expectations
  • Where to start
  • How to deal with an “Automation Project”
  • Sane approaches to “Automate ALL-THE-THINGS”
  • Dealing with unintended consequences
  • Recognizing the perception difference between product software “bugs” versus “incidents” created by automation
  • Entropy and you
  • Why Testing isn’t just the QA Team
  • Rise of the Machines

velocity_nyc_2014-headshotWaldo is a Geek, Problem Solver, Systems/Operations Engineer, and DevOps Enthusiast.  He tries to use his Viking Privilege superpowers for good, but is a flawed human being.  He aspires to be a better bio writer one day.

2016 Speaker Proposal: Tyranny of the Minimum Viable Product (MVP)


We, as an industry, do ourselves a disservice by claiming that a product is ‘MVP’, short for “Minimum Viable Product”.  Instead of labelling them a “Proof of Concept” or “Experiment”, many such projects fail at the tests of being “Viable” or “Product”.  Let’s be more humane to our peers and customers by labelling “Experiments” as such, and ensuring that Products are minimally Production-ready.


While continuously iterating on a product is absolutely important to ensure that you’re working on products that people want, calling them “MVP” does not give an excuse for not making them Operable.  There are terms for projects that are not Viable Products; Let us encourage the use of “Experiments” or “Proof of Concept” for such things that we want to run, but are unwilling to apply Production-ready rigor to, so as to set more reasonable expectations of support levels.
In this talk, I will discuss the points that make a Product “Minimally Viable”, such as:

  • Automatically Deployable
  • Config-managed, where applicable
  • Basically Monitorable (app health checks), including Pager decisions.
  • Shipping Metrics
  • Shipping Logs
  • Scalable
  • Reasonably Documented

velocity_nyc_2014-headshotWaldo is a Geek, Problem Solver, Systems/Operations Engineer, and DevOps Enthusiast.  He tries to use his Viking Privilege superpowers for good, but is a flawed human being.  He aspires to be a better bio writer one day.

2016 Speaker Proposal: Watch Everything, Watch Anything: Anomaly Detection using InfluxData’s Kapacitor by Nathaniel Cook

Five years ago Ian Malpass posted his “Measure Anything, Measure Everything” article that introduced StatsD to the world. Since then DevOps has grown and defined itself around the ideal to measure everything.

Now it’s time to take it further. “Watch Everything, Watch Anything”. At any DevOps shop you quickly realize that you can’t watch everything even if you put it on big screens throughout the office where everyone can see. You need something to watch the data for you and to tell you what is important and what is not. The open source Kapacitor project, the last piece of InfluxData’s TICK stackis just that tool.  Kapacitor can watch all your metrics looking for anomalies and notify you of interesting or important events in your metrics. Today‘s presentation will walk you through the basics of anomaly detection with Kapacitor, how it works and how to know which algorithms to use for your various metrics.

Nathaniel CookI graduated from BYU with a BS in Physics and a minor in CS. I am attending UIUC getting a masters in CS online. I worked for Qualtrics for the past 6 years, started as a developer and then moved into the DevOps role a year or so after starting. I started at InfluxData in Sep of 2015 and have been building Kapacitor ever since. I am a hobbyist and love home automation with several raspberry pies controlling various aspects of my home. Most recently I got into 3D printing one of my first projects was to print sections of the Minecraft world I have hosted for friends and family over the last year.

Church and State

Salt Lake City DevOps Days Official Announcement! June 14th and 15th, 2016

Church and State

We are announcing Salt Lake City DevOps Days: June 14th & 15th 2016 at Church and State. This conference It will be a two day event with conference talks in the morning, openspaces each afternoon and an evening social gathering. Tickets are available immediately for $99 for the first 100 people to register.

DevOps is “a cross-disciplinary community of practice dedicated to the study of building, evolving and operating rapidly-changing resilient systems at scale.” — Jez Humble of Chef.

You can contribute to SLC DevOps Days one of the following ways:

  1. Propose a session for the conference talks: Your talk will be a 45 minute session where you can speak about the next big things happening in world of DevOps. For a detailed description, please see our call for proposals.
  2. Propose a lightning talk: If you have a strong DevOps related message that you really want to share with the DevOps world, we encourage you to send us a proposal for an Ignite presentation. Ignite presentations are  5 minute talks with 20 slides which auto-advance. The only rule is don’t pitch products.
  3. Register and show up: One of the success factors of previous devopsdays is the openspace-format: it fosters the sharing of ideas and insights in good old fashioned, non-technical way by just talking to each other. Last time it even got people who would normally never talk, into the conversation.
  4. Sponsor the event: Registration fees only cover part of the budget. Therefore we actively seek sponsors to make this event happen. In return we hope to make it valuable for you to sponsor Salt Lake City DevOps Days.
  5. Volunteer: We’re expecting 300 people to show up to this event. At this time, we’re asking that you volunteer to help make this event a success. We’re going to need Event Assistant, Floater, Guide, Registration Assistant and Ushers. You’ll be the envy of your team and will get some free swag for your time.
  6. Tweet, Post and e-mail: Help us get the word out by emailing everyone on your team. Share our website, https://slcdevopsdays.org, Tweet about it using the hash tag #slcdevopsdays and post this to your friends on Facebook https://www.facebook.com/slcdevopsdays.

Early bird tickets are available immediately for $99 until April 15th, 2016. Hurry, only 100 are made available at this price. Once they’re gone, tickets are $149.

Please feel free to give us feedback by contacting the organizers.


February 2016
« Jan   Mar »