IDI2020 - Incontro DevOps Italia
21 - 22 Oct 2020 at Internet

19

If you want to know what really drives the world, just follow the money. This talk will offer a different view on common cultural mantras and technical trends of the Internet, such as "the Internet should not be regulated", "there should not be national borders online", "let's encrypt everything" and "the Web is the Internet", showing how in today's economy they suit the business and political interests of the big Internet giants and of surveillance capitalism in general. It will start by looking at where the revenues of these companies actually come from, and show how regulation and personal/national points of network control, like ad blockers and content filters, represent a big strategic risk for these companies, explaining their efforts to make those practices culturally unpopular. As a case study, the new DNS-over-HTTPS protocol will be examined in technical and policy terms, showing how, while providing better privacy and security in some use cases, it also addresses that strategic risk for those companies. To end on a positive note, regulatory efforts for a European political response will be summarized, and the open, cooperative, interoperable model based on free software and open standards will be presented as the possible way forward for the European Internet community.

11

Challenges and risks during adoption, and what to look for beyond. With microservices becoming mainstream a few years ago, many organizations have now adopted them; even though all are paying the price in terms of training, solution complexity and operational costs, few are reaping the promised benefits. Lower velocity, quality and performance issues, along with an overall lack of visibility are what we hear about most often. In this session, working from our experience as advisors to software development teams, we’ll walk you through some of the symptoms you might experience, their possible causes and some potential solutions. Furthermore we outline a pragmatic adoption framework to guide the technology choices in terms of your specific domain, goals and constraints.

Kubernetes is one of the most famous and used open-source projects worldwide and in all types of projects. The recommendations while deploying it are really specific: Create it with three or five master nodes to stay in HA (High Availability). Not all have to be configured in high availability. We will review considerations, recommendations, and we will help you to make the right decision when creating these trending infrastructures. We will experiment with destroying the master node of a single master Kubernetes cluster having a running application on it.

7

PCI-DSS è un acronimo che sta per “Payment Card Industry - Data Security Standard” ed in UK rappresenta il modello legale da soddisfare se si decide di gestire autonomamente il proprio sistema di pagamenti online. In questo talk racconterò le strategie che io ed il mio team abbiamo messo a punto per realizzare uno dei primi sistemi di pagamento, certificati PCI-DSS di primo livello, hostato interamente in AWS public cloud. Racconterò di come Dev ed Ops abbiano lavorato assieme, descriverò l’architettura dell’infrastruttura, il tipo di automazioni e quali gli accorgimenti utilizzati per la realizzazione della piattaforma.

Si può fare DevOps in una banca commerciale tradizionale? Nel talk illustreremo l'approccio che abbiamo seguito per introdurre i primi elementi di DevOps mostrando come è possibile cambiare anche in un mondo caratterizzato da forti vincoli normativi, software legacy e un contesto culturale di stampo tradizionale. Discuteremo anche cosa stiamo imparando durante questo viaggio e come il percorso sta evolvendo man mano che lo percorriamo. In particolare stiamo vedendo davvero ""sulla nostra pelle"" quanto il DevOps sia un cambiamento culturale e di metodo e non solo di tecnologia.

2

Illustro il nostro template di sviluppo open-source per lo sviluppo, il test ed il delivery di progetti principalmente basati su Python, Django, PostgreSQL, uWSGI e React che usiamo in produzione di progetti di medie e grandi dimensioni, per fornire servizi web o mobile. In questo talk il pubblico potrà vedere come noi mettiamo in pratica le regole dello sviluppo agile in ambito web usando il linguaggio ed il framework web che preferiamo con l'utilizzo di tecnologie di orchestrazione e CI/CD il tutto racchiuso nel nostro template open-source e riutilizzabile da tutti così com'è o adattabile anche con altri linguaggi di programmazione o framework. In definitiva vorremo portare l'esempio della nostra azienda dove, pur non avendo una figura dedicata esclusivamente ai processi specifici del DevOps ne usiamo l'approccio coordinandoci tramite il nostro template.

3

Software Configuration Management (SCM) is an important foundation for any type of project to work in a smooth and successful way. So there is no reason that SCM shouldn't be useful and necessary also on DevOps projects However, all projects are not alike and there is no single “one size fits all” SCM that will work in all contexts and for all development methods. SCM has well-established practices for more traditional ways of development, but very little is known about what SCM is needed and how it should be carried out in a DevOps context. DevOps projects probably need SCM done in different ways than “traditional” projects - and operationally by different (non-SCM) people. The difficult thing is not the concepts and principles of SCM - but to figure out what part of them can be helpful in different situations and in which way they should be implemented. As SCM experts - with a DevOps interest - we decided to investigate more closely how to match up SCM with DevOps. There is already some knowledge about how to do SCM in DevOps contexts, but it is fragmented and scattered around - and it is incomplete. We have researched “prior art” to collect existing wisdom. We have also analysed DevOps and SCM to identify areas where we could complete the existing knowledge. In this presentation, we will present and discuss our findings so far.

Cortex is a scalable, highly available, multi-tenant, long term storage for Prometheus. It is designed to reliably ingest and query millions of time series per second, while keeping the operational cost low. But if you try to get started, it is easy to get lost in the tens of different components and config options. In this talk, we will introduce Cortex and its architecture in an approachable way. We’ll start from the minimum setup required and we’ll iteratively cover the Cortex core design principles, sharding, replication, failure modes and a series of techniques employed by Cortex for accelerating PromQL queries. Finally, we’ll look at what the future holds for us and the upcoming features and improvements in Cortex. Cortex is an open source project, released under the Apache 2.0 license, and in the CNCF sandbox stage.

Hear how Microsoft embarked on their own DevOps journey reducing a three-year release cycle down to three weeks. This required changes to their people, process and products. April Edwards is a Senior Software Engineer and will share insights into the agile transformation of Microsoft's Developer Division and showcase newest additions to Azure and how to use Azure DevOps to deploy any language on any platform, including Windows, Linux, mobile and containers.

Terraform is a beautiful tool we all love; it is multi-cloud, meaning you can write code to provision resources from all major and many minor cloud providers. Sadly the code you write is vendor specific as resource descriptions are not portable between providers. What if your boss or customer wants a cloud-agnostic solution? Do you need an external tool or Terraform language is rich enough to allow cloud-agnostic code? I stand that 0.12 is so powerful that you can create provider-agnostic modules and write code that works the same in GCP, AWS and Azure.

5

Containers have been a driving force in this industry for the last 5+ years. In the meanwhile we have seen the raise of other compute patterns, such as serverless. 2020 seems to be the year where the line between containers and serverless starts to blurry. We are seeing the raise of container serverless platforms (e.g. AWS Fargate) as well as the raise of higher order abstractions above container platforms (e.g. OpenFaaS, ECS CLI v2, …) that allows developers to focus on their code instead of managing containers. In this session we will discuss how the serverless benefits are starting to permeate into the container ecosystem and we will provide real life examples of how some AWS and OSS technologies can be used to abstract and remove part of the undifferentiated heavy lifting developers often need to take care of.

4

While it's common wisdom that testing in production is a bad thing, it's actually a fundamental part of the work of most deployments in a fast-paced environment. In this talk, starting from my experience managing difficult transitions/rollouts for Wikipedia, I will underline what and how is worth testing in production, and some recommended patterns for doing so, both at large and small scale.

5

Dopo aver adottato gli orchestratori di container, i microservizi e aver industrializzato i ltutto tramite le pipeline CI/CD, il prossimo collo di bottiglia potrebbe essere la pubblicazione dei servizi ed il controllo delle loro comunicazioni. Per superare questi limiti potrebbe essere necessario introdurre in azienda una Service Mesh. La Service Mesh offre osservabilità, resilienza, controllo del traffico e sicurezza per le comunicazioni tra servizi distribuiti, aggiungendo un disaccoppiamento tra il codice applicativo(Dev) e le preoccupazioni a livello di servizio e gestione delle sessioni (Ops)

This "talkshop" introduces the concept of Kubernetes Operators and guides the attendee in writing a simple one on OpenShift using the Ansible automation framework. Level: Intermediate, for developers and system admins, no programming knowledge needed.