Adventures in DevOps

Latest Episodes:

DevOps 017: Improving User Experience by Logging with Grant Schofield Dec 10, 2019

Show notes

In this episode Adventures in DevOps, the panel interviews Grant Schofield. Grant is Director of Infrastructure at Humio. He being by discussing the growth of logging and logging tools. Grant explains the business value of logging and analytics. He shares some real-life examples of how longing helped gain insight into the user experience. The panel wonders how Humio takes the data gathered in the logs and separate out specifics of user experience. Grant explains that by aggregating all data in one place Humio uses the logs, tracing and other metrics to draw conclusions about user experience. He shares some of the conclusions that can be drawn from that data and explains that the conclusions all depend on what you are looking for. The panel discusses how tracing traditionally works and asks Grant what process Humio uses to good sampling. Grant explains that sampling is a good way to save on costs and depends on how much indexing is taking place. He explains that knowing when to sample is very important if you want an accurate sample. Compliance concerns are the next topic the panel discusses with Grant. He explains what Humio does to remain compliant and keep user info safe and private. The panel moves on to discuss index-free logging. Grant explains how index-free logging works. He explains how fast it is and how easily clients can retrieve their data. Panelists

Nell Shamrell-Harrington
Scott Nixon
Charles Max Wood

Guest

Grant Schofield

DevOps 016: Kubernetes as Infrastructure Abstraction with Oleg Chunikin Nov 26, 2019

Show notes

In this episode of Adventures in DevOps the panel interviews Oleg Chuninkin, CTO of Kublr. Oleg starts by explains what Kublr is all about and how he got the idea of using Kubernetes as an infrastructure abstraction. He and the Kublr team were trying to decide the most productive way to think of Kubernetes. Oleg advocates for using Kubernetes locally and shares how you can then orchestrate your architecture so you can see what it will do in productions. Charles breaks down a few of the ideas Oleg shares. Oleg explains how the portability of Kubernetes can be used and shares recommendations with the panel on how to run a Kubernetes in a lightweight way. The panel asks Oleg about the pressure for a cloud independent service and how these effects application requirements. Oleg shares some resources in answer. Moving on the panel considers Olegs comment about the layered architectural approach. Oleg outlines the layered architectural approach and explains what he means by layered. He explains the benefits of this approach. Panelists

Scott Nixon
Charles Max Wood

Guest

Oleg Chunikin

The MaxCoders Guide To Finding Your Dream Developer Job Nov 20, 2019

Show notes

"The MaxCoders Guide to Finding Your Dream Developer Job" by Charles Max Wood is available on Amazon. Get your copy here today only for $2.99!

DevOps 015: Shiplane with John Epperson Nov 05, 2019

Show notes

This episode of Adventures in DevOps is joined by John Epperson. John is a developer and has worked in DevOps for his whole career spanning about 12 years. He is also the author of Shiplane. John made Shiplane after working with Docker for a while and getting fed up with some features not being as he wanted them to be. The panelists begin the discussion with going over John's talk “The Good, The Bad, and The Ugly”, where he covers three camps of developers namely pioneers, settlers, and townsfolk and the panelists go into detail about these camps. They also talk about how docker and Shiplane fit into this idea. John shares that Shiplane is a tool that converts docker-compose yaml files into production-ready docker deployments. John details more of how shiplane came about, the problems it solves, how it works, and how it helps users to cross knowledge barriers. He also shares how Shiplane provides customizability as it will give the user everything they need but lets them choose the pieces they want. The panelists also discuss what websites are running Shiplane as well as some of the support that Shiplane has. Currently it does not support Kubernetes but John is working on it. Chuck then asks John what the use case is for running Shiplane. They discuss specifically whether Shiplane is used locally, or as a SAAS service and why it is the way it is. Nell then asks John what his hopes are for the future of Shiplane. John would like to have Kubernetes capability added within a year. If someone wants to get involved with Shiplane they can do so by trying it out, reaching out to him on his discord server, and helping him find edge cases. Panelists

Nell Shamrell-Harrington
Scott Nixon
Charles Max Wood

Guest

John Epperson

DevOps 014: Continuous Delivery With Julian Fahrer Oct 22, 2019

Show notes

Julian Fahrer is a software engineer with a systems administration and operations background. He currently works at Hover and is helping them move towards continuous delivery. Nell opens up the discussion by asking Julian to explain what continuous integration and continuous delivery are. He shares that continuous integration revolves around having testing and automation around the code being pushed to ensure that it works and conforms to standards. Continuous delivery feeds off of the concept of continuous integration and is the ability to deploy to any environment at any point in time. Chuck puts these ideas together by saying that continuous delivery and continuous deployment is about making sure that it's possible to deploy at any time and actually doing so. The next topic covered by the Adventures in DevOps panelists is how to handle apprehension around having a continuous integration / continuous deployment (CI/CD) pipeline. Julian shares that big cultural changes are required to make CI/CD successful. They share that a lot of organizations are nervous about code with breaking changes immediately being deployed. One way to handle this vulnerability is by hiding features behind feature flags so that only certain people, for example the QA team, will have access to the feature.They share why an organization would want to change to a continuous delivery pipeline and some real world examples that they have experiencedChuck asks what some prerequisites are for a CI/CD pipeline. Scott says that one thing that needs to happen is backfilling areas of the application that previously did not have tests and expanding the testing coverage of the system. Julian shares that these prerequisites depend on how changes are made and shares some specific examples of what that entails. Nell asks what it takes for an organization to be able to do continuous delivery. Julian shares that it is mostly process driven. They establish some rules such as shortening the lifetime of development branches and improving how they are deployed. They also have a discussion on how integration tests should come about and who should write them. Julian mentions that you want to empower people and give them the tools they need to succeed. They then cover some of the work that Julian has done with Hover and some of the details of the continuous delivery environment he is building and the steps they took to begin moving towards that workflow.The topic then moves to dependency management. Nell asks Julian how he approached dependencies in his applications. Julian details how his usage of containers and specific tools helped him. For external environments he says that a database is required and that it would help to have standards for managing dependencies. They share how the current development culture is to give the QA team enough time to test a feature before it goes out. With a CI/CD environment, feature flags can be used to gradually roll out a changes and if a certain users needs a specific set of features then an individual environment can be spun up for their use case. The panelists share some thoughts on environments setup and production best practices and tooling. Panelists

Nell Shamrell-Harrington
Charles Max Wood
Scott Nixon

Guest

Julian Fahrer

DevOps 013: Application Monitoring Using RED With Dave McAllister Oct 15, 2019

Show notes

This episode of Adventures in Devops features Dave McAllister. Dave has an extensive background in open source starting in 1994 working with early versions of Linux. He thrives on the concepts of emerging technologies and being able to innovate things. He also loves understanding what people are doing with emerging technology. The discussion opens up by introducing the topic of multi-dimensional monitoring in RED. Dave gives us an introduction into RED as a subset of google's SRE Golden signals. RED stands for rate, errors, and durations and is a concept that is designed for working with micro services. The DevOps panelists discuss concepts such as saturation and how to ensure correct results from their micro services using the RED concepts as well as some best practices for managing micro services. Nell asks about the scope of RED and whether it works with the big picture of what the micro service is doing. Dave shares that the scope of RED pertains to both. RED helps with observability and how to get the right signals out of all the noise and how to respond once the correct signals are found. He shares that RED should be a set of metrics in a dashboard that can be aggregated. He explains that RED gives the user a way of grouping data together and helping them to normalize functionality and find trends. The next topic covered by the DevOps experts is how to map the metrics seen in RED to the user experience. Dave explains how RED monitors the users activity and can put together metrics based on what they're doing. Using RED to follow user metrics will help to identify trends in where users will have issues and identify problem areas. Using micro services with RED introduces a level of granularity that can be monitored to help improve the performance of the application and improve scaling. RED helps with these improvements most notably by improving reaction time once a problem is found to help correct it as soon as possible. The panelists discuss some real world examples and how real world activities and human tendencies can alter patterns seen in the monitoring. Dave points out that one of the strongest recommendations he can make about RED is its ability to start simple and scale upwards as needed. The panelists then go on to discuss the human aspect of RED, how a team would react to changing, and how RED really requires a true DevOps team to reach its full potential. The panelists then share experiences they faced earlier in their careers as developers and how RED could have helped them. Nell brings up the idea of service meshes and how RED applies to them. Dave starts by introducing some problems in micro services and service meshes and the opportunity that exists for RED to come in and help solve those problems. He explains how service meshes in micro services give you duration that you don't have to implement. They finish with covering the usage of Kubernetes operators. Panelists

Nell Shamrell-Harrington

Guest

Dave McAllister

DevOps 012: Containerizing an Application Oct 08, 2019

Show notes

On this episode of Adventures in DevOps the panelists talk about how to containerize an app in a container using services such as Docker or Kubernetes. Charles mentions that he is wanting to host a Ruby on Rails application and Scott has experience with web deployments in Python and Django. They share that regardless of the languages and technologies used, the principles behind deploying a containerized application are the same.The next topic covered by the DevOps panelists is how to create containerized instances using Docker. The discussion centers on creating the container instances on a network and best practices that can be used for managing and maintaining them. The DevOps panelists cover the difference between an internal network deployment and what is needed to push a containerized application to a cloud instance using Kubernetes, Docker, DigitalOcean, or another service. Scott asks Charles if he is using a good CI/CD (continuous integration/ continuous delivery) mechanism, to which Charles shares that he is using Cloud 66 for its ease of use. When it comes to storage, they discuss whether to use the cloud services storage option or using a container to run a Postgres engine. They also go over the best way to run the shell commands for configuring their environment.The topic then moves to security and asset management when using Docker. Charles mentions that he tries to avoid putting API keys or other secrets into the Docker file and Scott shares how he uses a secrets management platform such as AWS Parameter Store, AWS Secrets Management, and others. They discuss the benefits that are provided by secrets management platforms. The DevOps panelists also talked about how a CI/CD platform can help with the portability of a containerized applicationCharles asks Scott how to begin with helping someone who wants to take their existing application and containerize it using the Docker approach. Scott responds with describing how he would look at where there application is and help them configure their containerized instance and set it up on a container service such as Docker or Kubernetes. Scott mentions that the strength of moving an application to a container service helps the application to scale up and down as needed. The container model helps a team to locally replicate their CI/CD model and be able to test it locally before the application is pushed to the cloud instance.The panelists talk about reverse proxies for pushing a Ruby on Rails app. Chuck prefers nginx and Scott emphasizes that it is important when pushing to production to spend some time verifying that the reverse proxy configuration is correct and won't expose the application in a way that it shouldn't be. Chuck shares that the use of a cloud platform such as GCP (Google Cloud Platform) of Microsoft Azure reduces the overhead with configuring containers and building images so that when he hands it off to the platform, it will take care of a large part of the setup including creation of the reverse proxy. Panelists

Charles Max Wood
Scott Nixon

DevOps 011: Let's Talk Agile Sep 24, 2019

Show notes

In this episode of the Adventures in DevOps podcast, the panelists talk about Agile. Nell starts the discussion by mentioning that Agile came out of the Scrum movement and gives a brief history of how it started. She reads out the Agile Manifesto, and Scott explains what it is. They agree that “move faster slowly” can be considered as a motto of Agile. Scott talks about the testing aspect of it from a philosophical perspective, and Nell chimes in with comparison to how things were being done in the past versus now.They discuss the 12 principles behind the Agile manifesto. One - Our highest priority is to satisfy the customer through early and continuous delivery of valuable software. Two - Welcome changing requirements, even late in development. Agile processes harness change for the customer's competitive advantage, basically valuing learning and improvements. Three - deliver working software frequently, from a couple of weeks to a couple of months with a preference for a shorter time scale, the focus being on the code being in a working state. Four - Business people and developers should collaborate daily through the project. Scott elaborates on the point further, and shares his experience where it was challenging to work with the business side of the project.Five - Build projects around motivated individuals, give them the support they need and trust them to get the job done, trust being the most important factor here. Six - The most efficient method of conveying information is having a face-to-face conversation, here the panelists provide a counter-argument that in the era of online platforms and remote work, this is not necessarily essential. Seven - Working software is the primary measure of progress. Eight - Agile processes promote sustainable development. The sponsors, developers, and users should be able to maintain a constant pace indefinitely. ‘Sustainable' is key here, avoiding overworking and burnout is important anyday. Nine - Continuous attention to technical excellence and good design enhances agility. This means doing the full set of code reviews and tests and still deploy software quickly and making sure that it doesn't break things, if it does, it has to be rolled back.Ten - Simplicity - the art of maximizing the amount of work not done - is essential. This highlights the need for focus in the product, and not writing a lot of code that could eventually be unusable. Eleven - The best architectures, requirements, and designs emerge from self-organizing teams. Panelists talk about people involved in open source projects as an example of such self-organizing teams, where they work on their own schedule and have great interest in what they are building. Twelve - At regular intervals, the team reflects on how to become more effective, then tunes and adjusts its behavior accordingly. Retrospectives, reflecting where you are, how the team works together and documenting results as some sort of a learning exercise, are some of the practical examples of this.Nell then talks about the three layers of Agile infrastructure by Patrick Debois - considered as the father of DevOps. These constitute of technical - the hardware and software, project - the process that introduces changes in the environment, and the operations - the process of keeping the environment working.They state that we can build effective technology if we can relate to each other as humans, and discuss how Agile is perceived in certain controlled environments, if it's possible in an Agile fashion and still have prescriptive solutions. They end the show with picks.Panel

Nell Shamrell-Harrington
Scott Nixon

DevOps 010: What Makes a 10x Engineer Sep 17, 2019

Show notes

Episode Summary
On this week's Adventures in DevOps, Chuck talks about the recent Twitter thread about 10x engineers. He goes through each of the points in the tweet and talks about each of them in turn. There are only two points he sort of agrees with, and believes the rest to be absolute garbage. One of the issues with this tweet is that it doesn't define what a 10x engineer is.Defining a 10x engineer is difficult because it is also impossible to measure a truly average engineer because there are many factors that play into measuring productivity. Chuck turns the discussion to what a 10x engineer is to him and how to find one. A 10x engineer is dependent on the organization that they are a part of, because they are not simply found, they are made.When a 10x engineer is added to a team, the productivity of the entire team increases. Employers have to consider firstly what you need in your team and how a person would fit in. You want to avoid changing the entire culture of your organization. Consider also that a 10x engineer may be hired as a 2x engineer, but it is the employer that turns them into a 10x engineer.Overall, Chuck believes these tweets are asinine because it's impossible to measure what makes a 10x engineer in the first place, and hiring a person that fits the attributes in the list would be toxic to your company.Panel

Charles Max Wood

DevOps 009: Learning DevOps Sep 10, 2019

Show notes

Episode Summary
In this episode of the Adventures in DevOps podcast, the panelists talk about getting started with DevOps and learning its principles and technologies. Lee and Scott mention that they share a similar journey, initially starting as system administrators and working their way into DevOps. Nell says that she was a Ruby and .NET developer before learning cloud technologies such as Heroku, Chef, and Docker.To someone with a programming background wanting to get into DevOps, Lee highly recommends three books - The Practice of System and Network Administration, The Practice of Cloud System Administration: DevOps and SRE Practices for Web Services - Volume 2, and Time Management for System Administrators. He mentions that they are technology-agnostic and help in understanding what is really going on in the system. Scott stresses on the importance of understanding basic networking, linux systems and commands, bash scripting and knowing the core fundamentals and interactions of the underlying systems. He suggests going for online resources, tutorials and boot camps which are plenty nowadays and easily accessible as well.Nell advises listeners to pick a cloud provider such as Digital Ocean and learn the basics by working with it, which can later be applied to any other cloud provider. She also recommends learning programming languages to get a good software development foundation. The panelists talk about working on specific projects and getting hands dirty rather than traditional learning, to enhance their DevOps skills.Nell explains how they implement the 'learning by doing' concept at work. She mentions that understanding how virtual machines and physical servers work is crucial before moving on to learning about containers. Lee suggests Vagrant - a tool for building and managing virtual machine environments, as a good resource for the same. They caution that while it is important to learn things during our time off, self-care and setting healthy boundaries is paramount. It is ok to take longer to learn something given that DevOps as a whole can be extremely complex, and working in a non-pressure environment can be very beneficial. They recommend Ops School for beginners which is a comprehensive program aimed at people who want to get into operations engineering.They then share their insights on monitoring and how to get started with it. Lee recommends the book Practical Monitoring, and Sensu - a monitoring event pipeline, as good starting points.DevOps can be considered as a technical as well as a cultural movement. To that effect, they discuss where should people start learning the associated cultural elements. Nell recommends reading the books - Effective DevOps, and The Phoenix Project and Crucial Conversations. Lee advocates going old school with the writings of Richard Stallman, and books of the 80s and 90s such as The Art of C Programming. They mention that conversational skills and dealing with people are critical skills in today's work environments. Scott recommends Google's Site Reliability Engineering books which have a lot of great stuff to build a solid foundation and are also free to read online.Finally, they talk about how to keep learning and expanding knowledge. Some effective suggestions discussed include extensive practicing, working professionally, solving business problems, building expertise in programming, and attending DevOps Days events and Linux Users groups. They end the episode with picks.Panel

Nell Shamrell-Harrington
Lee Whalen
Scott Nixon

Our TOPPODCAST Picks

Follow Us

Stay Connected

Related Podcasts

Links

Stay Connected