Anonymous DevOps Club
Every Tuesday in the evenings a club of anonymous DevOps fans gathers. Our tasks are much more ambitious than just sharing your problem and, perhaps, getting a sovset. We discuss all the industry trends so that it would be nice to come to the conference and the DevOps section ourselves.
Since July, we have been working on DevOpsConf Russia , a professional conference on the integration of development, testing and operating processes, which grew out of RootConf and will be held on October 1 and 2 in Moscow, in Infospace.
Today I will tell you how we do it, what reports we are trying to select, and ask the speakers.
The focus of the conference is integrity . We will talk about DevOps from different angles - from the side of development, operation, processes, management and tools. That is why we formulated such a set of sections:
In our opinion, this set fully describes all the important aspects. A program covering all these issues will be comprehensively useful. On the other hand, the main task is to maintain balance , not to start talking about processes and culture in isolation from technologies and tools. Therefore, we ask our speakers to restore the process in the company, clarify goals and objectives and link them to technological solutions.
I will explain with an example: a super-speaker of all DevOps conferences is going to be Alexey Vakhov , CTO in the company Uchi.ru.
We talked with Alexey and found out that since his last report, there have not been any special technological innovations in their infrastructure and supply line. After discussion and a little brainstorming, we just decided to go from the process. How exactly does an artifact with code get from the developer to production, what happens to the artifact, how are the guys protected from bugs and bugs throughout the entire pipeline? Alexey agreed that this was a really fresh look at what was happening inside the company and decided to redo the report in this vein. And now we, as listeners, will not only be able to learn about the experience of using various tools in the company of Uchi.ru, but also find out how the process of continuous delivery works.
This is just one example of the joint work of the Program Committee and the speaker on a report. Every application goes through such a collective discussion, we delve into the details so that everything works out perfectly. Therefore, now I will tell you about all the reports, but in each topic I will choose one or two very interesting ones.
Infrastructure platform
They often talk about kubernetes now, and okmeter.io concluded that they also need k8s in production. Although they do not even have CI / CD, but there are limitations such as the required maximum resiliency and the lack of human resources for this task.
In his report, Nikolai promises to tell you how you managed to solve this problem:
Experience okmeter.io will be useful if you are looking for a balance entry threshold / benefits k8s / control / performance.
From this report we will find out: what data on the application's work in Booking.com is collected for monitoring; how they are aggregated; what metrics and logs are stored; as analyzed.
Consider the popular opensource monitoring solutions , analyze the limitations and features, let's see what tools had to be implemented independently.
Continuous delivery
Ansible is one of the popular configuration management systems, has a low entry threshold, conveniently expands with third-party modules, allows code reuse , and has several other advantages. But the introduction of a configuration management system in the forehead helps only at first. After some time, it becomes quite difficult to maintain an expanded number of roles. Alexander Kharkevich from EPAM Systems in his report will talk about the mechanism of continuous supply of roles , as the most effective way to support them. Let's analyze the development of both public roles and public roles, but with test runs in private infrastructure.
All collect a variety of technical indicators and some business metrics: revenue, retention, quality . Unfortunately, very often these metrics are analyzed separately from each other, and no one tries to relate them. Do you know how much money a web server brings to you? How to see problems when all technical monitoring systems are green? How much money does a business lose when the database is 90% loaded? And 50%? Come to the report of Basil - we will understand.
This report will be submitted from the perspective of a proactive devops-engineer and developer. Consider the entire life cycle of a (micro) service using a specific example of a python service:
Maxim will share experience that will allow administrators to delegate configuration and deployment of applications to developers with peace of mind . And developers should concentrate on features without significant downtime due to switching to devops activity.
SRE practices
All major cloud providers, from AWS to Google Cloud, provide various options for running MySQL or MySQL-compatible databases in the cloud. You can deploy virtual machines and set up your own cluster or rely on services and manage databases with the click of a button .
In a report by the chief architect of the cloud platform Funambol Renato Losio, we consider the operating costs of running relational databases in the cloud and how to integrate into the infrastructure as a code. Let us see if the time has come for serverless databases.
From this report, we learn not only about the accepted practices of SRE, Igor will illustrate his story with examples and the main promise is to give the most valuable thing - the generated service template , which includes the experience of several years of work.
Consider, in particular, such questions: service generation with the choice of the necessary modules; methods for testing and validating code; delivering and executing code in a container locally and in a cluster; structure of service modules from the best practices; polite shutdown service (graceful shutdown). And much more from the invaluable experience developing in containers.
From the report of Sergey we learn about managing secrets in Avito using Hashicorp Vault, and using Puppet and Kubernetes . Also, Sergey promises to list which bumps are stuffed in a year and a half of use, and will share his thoughts on how to fix it.
Anton Isanin is the author of Alfa-Bank's DevOps transformation strategy. From his report, we learn what problems the company faces in the process of transformation. Let's talk about a specific technical implementation : how to write a user story; how to organize development and testing; how to implement warm; monitor work in industrial operation, etc. There are many nuances, and it is useful to take into account the experience of companies that have already experienced the DevOps transformation.
See you at the big and not anonymous meeting of DevOps fans on October 1 and 2! Just do not forget to book a ticket in advance, closer to the conference it will be more expensive.
Since July, we have been working on DevOpsConf Russia , a professional conference on the integration of development, testing and operating processes, which grew out of RootConf and will be held on October 1 and 2 in Moscow, in Infospace.
Today I will tell you how we do it, what reports we are trying to select, and ask the speakers.
The focus of the conference is integrity . We will talk about DevOps from different angles - from the side of development, operation, processes, management and tools. That is why we formulated such a set of sections:
- Infrastructure platform.
- Infrastructure as a code.
- Continuous delivery.
- Feedback.
- Architecture in DevOps, DevOps for CTO.
- SRE practices.
- Training and knowledge management.
- Security, DevSecOps.
- DevOps Transformation.
In our opinion, this set fully describes all the important aspects. A program covering all these issues will be comprehensively useful. On the other hand, the main task is to maintain balance , not to start talking about processes and culture in isolation from technologies and tools. Therefore, we ask our speakers to restore the process in the company, clarify goals and objectives and link them to technological solutions.
I will explain with an example: a super-speaker of all DevOps conferences is going to be Alexey Vakhov , CTO in the company Uchi.ru.
We talked with Alexey and found out that since his last report, there have not been any special technological innovations in their infrastructure and supply line. After discussion and a little brainstorming, we just decided to go from the process. How exactly does an artifact with code get from the developer to production, what happens to the artifact, how are the guys protected from bugs and bugs throughout the entire pipeline? Alexey agreed that this was a really fresh look at what was happening inside the company and decided to redo the report in this vein. And now we, as listeners, will not only be able to learn about the experience of using various tools in the company of Uchi.ru, but also find out how the process of continuous delivery works.
So, if it’s interesting to you, like me, it’s not too late to register for the DevOpsConf Russia conference !
This is just one example of the joint work of the Program Committee and the speaker on a report. Every application goes through such a collective discussion, we delve into the details so that everything works out perfectly. Therefore, now I will tell you about all the reports, but in each topic I will choose one or two very interesting ones.
Infrastructure platform
Nikolay Sivko. Kubernetes for those over 30
They often talk about kubernetes now, and okmeter.io concluded that they also need k8s in production. Although they do not even have CI / CD, but there are limitations such as the required maximum resiliency and the lack of human resources for this task.
In his report, Nikolai promises to tell you how you managed to solve this problem:
- HA k8s on bare metal of toothpicks and blue electrical tape.
- How to organize a network of k8s, provide connectivity k8s cluster and neighboring servers.
- Why is the k8s service network not used, but headless-services and envoy + dns discovery are implemented wherever possible.
Experience okmeter.io will be useful if you are looking for a balance entry threshold / benefits k8s / control / performance.
Anna Stepanyan. Monitoring Infrastructure on Booking.com
From this report we will find out: what data on the application's work in Booking.com is collected for monitoring; how they are aggregated; what metrics and logs are stored; as analyzed.
Consider the popular opensource monitoring solutions , analyze the limitations and features, let's see what tools had to be implemented independently.
Continuous delivery
Alexander Kharkevich. Effective development and maintenance of Ansible roles
Ansible is one of the popular configuration management systems, has a low entry threshold, conveniently expands with third-party modules, allows code reuse , and has several other advantages. But the introduction of a configuration management system in the forehead helps only at first. After some time, it becomes quite difficult to maintain an expanded number of roles. Alexander Kharkevich from EPAM Systems in his report will talk about the mechanism of continuous supply of roles , as the most effective way to support them. Let's analyze the development of both public roles and public roles, but with test runs in private infrastructure.
Feedback
Vasily Ozerov. Revenue Based Monitoring
All collect a variety of technical indicators and some business metrics: revenue, retention, quality . Unfortunately, very often these metrics are analyzed separately from each other, and no one tries to relate them. Do you know how much money a web server brings to you? How to see problems when all technical monitoring systems are green? How much money does a business lose when the database is 90% loaded? And 50%? Come to the report of Basil - we will understand.
Architecture in DevOps, DevOps for CTO
Maxim Vikharev. DevOps-saga about template microservice
This report will be submitted from the perspective of a proactive devops-engineer and developer. Consider the entire life cycle of a (micro) service using a specific example of a python service:
- what a typical (micro) service consists of and depends on in a modern web application;
- what architectural features and primitives kubernetes use;
- how a (micro) service should be integrated into the process of developing and delivering applications, in order not to write kilometer instructions and generally do nothing at all;
- how to dock applications with system services in multiple production environments, multiple test and debugging environments, and why;
- what arrangements and solutions are needed between the infrastructure team and the development team for comfortable and efficient interaction.
Maxim will share experience that will allow administrators to delegate configuration and deployment of applications to developers with peace of mind . And developers should concentrate on features without significant downtime due to switching to devops activity.
SRE practices
Renato Losio. MySQL on the cloud
All major cloud providers, from AWS to Google Cloud, provide various options for running MySQL or MySQL-compatible databases in the cloud. You can deploy virtual machines and set up your own cluster or rely on services and manage databases with the click of a button .
In a report by the chief architect of the cloud platform Funambol Renato Losio, we consider the operating costs of running relational databases in the cloud and how to integrate into the infrastructure as a code. Let us see if the time has come for serverless databases.
Igor Dolzhikov. How to satisfy a SRE or Go service in a container in 5 minutes
From this report, we learn not only about the accepted practices of SRE, Igor will illustrate his story with examples and the main promise is to give the most valuable thing - the generated service template , which includes the experience of several years of work.
Consider, in particular, such questions: service generation with the choice of the necessary modules; methods for testing and validating code; delivering and executing code in a container locally and in a cluster; structure of service modules from the best practices; polite shutdown service (graceful shutdown). And much more from the invaluable experience developing in containers.
Security, DevSecOps
Sergey Noskov. Manage secrets with Hashicorp Vault
From the report of Sergey we learn about managing secrets in Avito using Hashicorp Vault, and using Puppet and Kubernetes . Also, Sergey promises to list which bumps are stuffed in a year and a half of use, and will share his thoughts on how to fix it.
DevOps Transformation
Anton Isanin. DevOps in Alfa-Bank
Anton Isanin is the author of Alfa-Bank's DevOps transformation strategy. From his report, we learn what problems the company faces in the process of transformation. Let's talk about a specific technical implementation : how to write a user story; how to organize development and testing; how to implement warm; monitor work in industrial operation, etc. There are many nuances, and it is useful to take into account the experience of companies that have already experienced the DevOps transformation.
See you at the big and not anonymous meeting of DevOps fans on October 1 and 2! Just do not forget to book a ticket in advance, closer to the conference it will be more expensive.
Subscribe to Ontiko's thematic newsletter on DevOps to get program updates as soon as they appear. We try to make the letters useful and non-intrusive, send news conferences, transcripts of reports, fresh videos.
By the way, a separate video can be monitored on the YouTube channel - all videos for the past few years have been collected there and the list is constantly growing.