Devops

If you've worked with Kubernetes for a while you should have seen that version lifecycle that Kubernetes has is pretty aggressive. Kubernetes is being actively developed which means that new releases come out pretty frequently - and all of us who are using Kubernetes, be it a managed or a self-hosted distribution, must adapt and adjust. If not, we may risk running our workloads on unsupported version of Kubernetes with lacking security and functionality fixes. In addition, when the time comes to upgrade there's a big risk that the change gap will be too large from the Kubernetes version you're running and the one you want to upgrade to - this normally ends up in a complicated and unstable upgrade process which I wouldn't recommend to anyone, especially in a production environment…😑 ...

Following up on the recent blog post series about Continuous Delivery to AKS with Azure DevOps Environments: 🐱 Part 1 🐱 Part 2 I would like to share a simple but really useful way to re-use the same Azure DevOps Environment in multiple deployment stages across multiple pipelines. Normally if you want to deploy multiple applications to the same AKS cluster, you would specify a target Azure DevOps Environment by writing it's name in environment property of deployment stage of every application: ...

Use case for pipeline templates In the world of complex enterprise applications and distributed systems you may have a need to perform many more actions and validations as part of a build pipeline than before: build an application, execute multiple types of tests like unit tests and API tests, perform security validations like SCA, SAST, container image scanning and scanning of third-party dependencies, perform application packaging and deployment, etc. That's when it's worth considering to implement a multi-staged pipeline where you can run several jobs in parallel and control application flow with stages. Each stage may then have it's own set of checks and validations. You may even have multiple applications which have similar build tasks as part of the build pipeline - for instance, if you have multiple .NET Web API applications, it's very likely that build pipelines for those will be similar to some extent. ...

Introduction Sometimes things may go wrong, also during ugprade of the application that is deployed to a Kubernetes cluster with Helm. When something goes wrong, you fix it and re-try a deployment with the new fix included. But then deployment operation fails with following error message: Error: UPGRADE FAILED: another operation (install/upgrade/rollback) is in progress So, what does it mean and how can you fix it? Let’s find out! Controlled, manual fix This happens typically when Helm attempts to roll out a new revision of an application and then something goes wrong in the process, like a bug in an application itself or an issue inside the Kubernetes cluster, which causes the new deployment to never get completed. This faulty deployment becomes dangling, therefore preventing all the future deployments to be rolled out. You can easily check the latest deployment status by retrieving the history of application deployments with helm history. ...

Code reviews and pull requests are a totally normal practice these days if you’re working together with other developers. Even if you’re working solo you probably don’t want to commit to main branch directly if you have CI/CD in place. That’s when adding changes through pull requests and code reviews come into picture. For many organizations and development teams this can be a tough process for many reasons where urgency, time limitations, overload with other tasks play a huge role. Nevertheless we still want to ensure that the established policies are being followed in order to ensure that every single line of code we commit is of highest possible quality. And there are ways we can enforce policies of course. One of the ways are automated checks - this is a really important one since we want to avoid human error or a multi-tasking PR reviewer to miss out on a critical bug in the changeset. But it’s also about humans, about collaboration and making the code review process easier both for the one committing changes and the one reviewing the changes. And for that there are quite a few tools available! ...