Dimensional Data

A successful DevOps practice relies heavily on metrics. Here at GitLab, we use seven key DevOps metrics to measure engineering efficiency and productivity. Like many teams, we use industry standard metrics, but in some cases, we approach this data with a unique GitLab point of view. Here’s the first in a multipart look at the DevOps metrics we at GitLab think are most critical for success. Compare your metrics and results with ours, and let’s get a conversation started. Master pipeline stability It’s important to be able to measure the stability of the GitLab project’s master branch pipeline. This metric tells us how stable the main branch is, and ensures engineers are checking out code that’s in good shape. Merge trains are key to this effort. Our target percentage for master pipeline stability is above 95%. Review app deployment success rate At GitLab we take review apps seriously. We measure their success rate so we can understand the stability of our first deployed environment after code change. Review apps are spun up at MR submission. It’s important to monitor our review app successful deployments because it’s the first place where code is integrated and deployed as one unit. This metric ensures the codebase can be installed, tested, and made available for the team to preview their changes before merging into the main master branch. Our target for review application deployment success is above 99%. Time to First Failure Time to First Failure (TtFF, pronounced as “teuf”) measures how fast we are providing feedback to engineers. This metric examines how long it takes from pipeline creation to the first actionable failed build. The idea is that if the commit is going to fail, it should fail fast and the fail signal should get to the engineers as quickly as possible. The shorter the time to first failure, the faster the feedback loop, and faster time to action to address those failures. Our TtFF target is less than 15 minutes. Open S1 bug age This metric focuses on the age of open S1 bugs. Many organizations measure time to close bugs. At Gitlab we focus on the age of bugs remaining. We structure the metric to focus on work that is remaining and can be actioned on. If we only measure time to close of fixed defects, we may miss addressing older defects and unintentionally incentivize closing of only newer defects. We like to look forward by asking ourselves “What’s left?” and “What can be done now?” rather than only looking backward at what’s already been done. Our target for S1 open bug age is under 100 days. Open S2 bug age This metric is similar to the open S1 bug age, but is focused on S2 bugs. Again, we measure the age of remaining open bugs rather than focusing on bugs that have been closed. Our target for the open S2 bug age metric is below 300 days. Merge request pipeline duration When a pipeline is started for a merge request, how long does it take to run? This metric focuses on the duration of merge request pipelines and its time efficiency. Within the total duration we break the data down into multiple stages The team then iterates and improves time efficiencies of each stage of the pipeline. This is a […]

If you work in an IT team of five people – or maybe you’re even a team of one – it’s easy to think your business is simply too small to use DevOps. But that’s not the case. A start-up or small and medium-sized business (SMB) is never too small to take advantage of a DevOps platform. In fact, DevOps is a great fit for a lot of SMBs, or small and medium-sized enterprises (SMEs). Here’s how to understand if it will work for your team or organization and how it could help grow your business in a competitive environment. The size of the business isn’t the issue Let’s be clear. If you are developing software, you need a DevOps platform. Size isn’t really the issue. No matter how small your business and your tech team, if you are iterating on software features, building applications, or automating parts of your product-related systems, then you do need DevOps. DevOps will even work for a team of one. Here’s how a DevOps platform can help an SMB: Start small to foster innovation One of the key aspects of DevOps is that it creates a collaborative atmosphere, even beyond the software and IT teams. Adopting a single, end-to-end DevOps platform when your company is small or your start-up is just getting off the ground will enable and encourage everyone – whether they’re in a technical role or work in accounting, sales or as a business manager – to all work together. And that will foster innovation by bringing in ideas from people in a range of demographics and business interests. And innovative ideas will help new businesses get a foot in the door and help all SMBs grow into more successful and bigger companies. Optimize your SMB for speed To get established in the market, start-ups and small businesses need to deliver compelling products quickly, and be able to efficiently support them. DevOps will enable your team to move from planning to production faster and with greater ease. A DevOps platform extends through the entire software development lifecycle, from planning all the way through to launching new features, conducting analysis, and gathering feedback. Simply put, DevOps will optimize your organization for speed, which is just what SMBs and SMEs need. Use DevOps to take on the “deep pockets” As an SMB, you likely don’t have the deep pockets and market penetration of your more-established competitors. How do you boost your odds when taking them on? One way to increase your competitiveness is to use DevOps to boost speed and efficiency as you create new products, new services, and new ways to communicate with your customers. When you can deploy innovative ideas faster than your competitors, you’ll have a definite advantage. Decrease your workload with automation When you have fewer hands to take on a huge workload, you need a way to not only speed production but to ease the number of tasks you’re facing – and all the headaches that come along with them. The automation that is part of a DevOps platform will mean less manual work when it comes to processes like design, testing, development, deployment, and monitoring. Automation helps small teams free up time to handle all the other projects on their to-do lists. Build security into software from the […]

It is possible to use GitLab as a best-in-class GitOps tool, and this blog post series is going to show you how. GitOps is an operational framework that takes DevOps best practices used for application development such as version control, collaboration, compliance, and CI/CD tooling, and applies them to infrastructure automation. This series of easy-to-follow tutorials will focus on different user problems, including provisioning, managing a base infrastructure, and deploying various third-party or custom applications on top of them, that can be solved by pairing GitOps with GitLab. Here are 8 tutorials on how to do GitOps with GitLab: 1. Here’s how to do GitOps with GitLab This tutorial sets the stage for what you will learn throughout the series, including the tech concepts you’ll need to know. 2. Infrastructure provisioning with GitLab and Terraform This tutorial walks you through setting up the underlying infrastructure using GitLab and Terraform. 3. Connect with a Kubernetes cluster This tutorial demonstrates how to connect a Kubernetes cluster with GitLab for pull- and push-based deployments and easy security integrations. 4. How to tackle secrets management This tutorial builds on the previous tutorial to show you how to use a Kubernetes cluster connection to manage secrets within a cluster. 5. The CI/CD tunnel This tutorial introduces you to CI/CD tunnels and shows step-by-step how to access a Kubernetes cluster using GitLab CI/CD. 6. Connecting GitLab with a Kubernetes cluster – Auto DevOps This tutorial looks at how you can use Auto DevOps with all its bells and whistles to easily manage deployments. 7. Connecting GitLab with a Kubernetes cluster for GitOps-style application delivery This tutorial shows you how to connect an application project to a manifest project for controlled, GitOps-style deployments. 8. Turn a GitLab agent for Kubernetes installation to manage itself This tutorial is the culmination of the previous tutorials and will teach you how to turn a GitLab agent for Kubernetes installation to manage itself. Read more about GitOps: “Want to learn how to pair GitOps with GitLab? Our eight-part in-depth tutorial series will take you step by step.” – Viktor Nagy Click to tweet