Kyiv DevOps Day (подія в архіві)

Mandi Walls, Chef (UK)
Transforming Enterprise teams to DevOps workflow:
Abstract: As large organizations become more interested in DevOps and the velocity it can offer, introducing new ways of working to teams with longtime habits and familiar workflows can be challenging. Shifting goals, new tools, and new skills create a stressful environment for technologists still trying to keep applications and services running. New tools should make work easier, not worse! This talk will cover some of the common pitfalls large organizations face when radically changing work as well as tips for technologists and managers for surviving the implementation of large changes.

Jakub Nabrdalik, Allegro (Poland)
Conway’s revenge: How ignoring org setup backfires in architecture, and vice versa"
Abstract: There are a lot of concepts in software development, that look perfect on the whiteboard, and then completely suck in practice. Due to years of things like implementing ESB where it’s not needed, programming with EJB, where it brings only tears and blood, our industry has build a distrust for Ph.D., ex-cathedra, design by committee and generally any theory not tested in reality. We have learned to make small steps, proof-of-concepts, spikes, proper research. Yet, even with small steps, TDD and the like, the architecture we design backfires months later in unpredictable ways. That happens quite often because architects disregard the organization’s structure and culture completely.

Alex Conway, DataRobot (USA)
Infrastructure Automation Tools
Abstract: At DataRobot we deal with automation challenges every day. This talk will give insight into how we use Python tools built around Ansible, Terraform, and Docker to solve real-world problems in infrastructure and automation.

Yegor Bugayenko, Teamed.io (USA)
Need Robust Software? Make It Fragile
Abstract: In any software project, the goal is to create something stable. We don’t want it to break in front of a user. We also don’t want our website to show an “internal application error” instead of a web page. We want our software to work, not fail. That’s a perfectly valid and logical desire, but in order to achieve that, we have to make our software as fragile as possible. This may sound counter-intuitive, but that’s the way it is. The more fragile your app is in development, the more robust it is in production. I will demonstrate, with practical Java examples, what Fail Fast means and how it helps us make software more stable.
See: www.yegor256.com/...​2015/08/25/fail-fast.html

Vsevolod Polyakov, Grammarly (UA)
Metrics: Where and How
Abstract: Nowadays it’s only a lazy one who haven’t written his own metric storage and aggregation system. I am lazy, and that’s why I have to choose what to use and how to use. I don’t want you to do the same job, so I decided to share my considerations concerning architectures and test results.

Marián Knotek, Forest technologies (UK)
ChatOps — making complex processes accessible.
Complex business processes can be automated with modern day automation and orchestration tools with relative ease. Those tools however are not built for use by end users and many such business processes cannot simply be scheduled, they require an event or interaction to be started. A good example would be a New Employee Onboarding process. This talk will introduce the concept of ChatOps, how it makes such complex automated processes easily accessible through instant message (IM) tools, and an example implementation.