Do it Like a Unicorn!
The new release of CloudBees Flow let's you deploy like Etsy and Netflix, with a click of a button!
In this software-driven world consumers are always expecting newer capabilities faster, which means Release teams need to deliver feature-rich software updates more often than the competition - and do that with ‘zero downtime’! This desire for ensuring ‘zero downtime’ and minimizing the risk of failed releases has given rise to many deployment strategies like Blue/Green, Canary, Dark Launch or Rolling Deployments. Until now, enterprises had to rely on heavy scripting, manual steps, weekend-long release meetings (and lots of praying) to maintain uptime as new versions of business critical applications were rolled out. The new version of CloudBees Flow changes all that. We continue to innovate, by extending CloudBees Flow's model-driven approach to deployment and pipeline orchestration, to also support advanced deployment patterns, such as “Rolling Deployments.” Instead of relying on scrips, customers can simply push a button and configure rolling deployment policies natively in CloudBees Flow, and then leverage them consistently release after release.
When we say 'push-button', we mean it!
Check out how this advanced deployment strategy looks like in the product itself:
Model your Rolling Deployment Strategy
There are several different rolling or phased-deployment models that teams typically pursue:
Some have very deterministic rolling deployment ‘phases’ where target machines are handpicked to be part of certain groups
Some want to roll with certain batch sizes - deploy to any 2 machines at a time until the entire environment is updated
Some want to do it based on ‘percentages’ – deploy in 4 batches of 25 of machines
Or some want to determine which machines to use per phase dynamically at run time!
Like everything else in CloudBees Flow, we give users total flexibility to decide their rolling deployment policies. Users can easily decide whether to use ‘Phase based Rolling ’ or ‘Batch based Rolling ’. You can then easily assign which machines belong to which group in Phase-based approach Or, choose a 'rolling' logic based on ‘batch size’ or of machines. It's that easy.
Process and Environment Fidelity
Our model-driven deployment has always allowed users to use the same parameterized process to deploy to different environments. This approach now extends to Rolling Deployments as well - allowing teams to use the same application process to deploy to different environments (QA, Pre-Prod, Prod, etc.) and choose whether or not to enable Rolling Deployments. This enables, for example, the Release team to deploy to UAT without rolling deployments, but then deploy to Production in a rolling fashion. And all this is possible at a click of a button. Since we recognize flexibility matters to support enterprise processes and large-scale deployments - at any stage users can easily choose to execute all phases in the rolling deployment, skip phases, change phase order or insert manual or automatic validations, if needed. Just because you want to move fast, you don’t have to increase risk, or sacrifice compliance and governance. And of course real time detailed view of every action is available to ensure progress and visibility.
There's More:
In addition to the breakthrough push-button Rolling Deployments capability, the new version of CloudBees Flow includes a number of other new features, such as:
Environment Reservation and Calendaring – increases visibility and reduces the chance for cross-team conflict during deployments through a new rich calendar view. This allows users to reserve environments for applications, pipelines and releases; define black-out/maintenance windows; and, offer policy-driven prioritization and conflict handling.
Automated Environment Discovery – reduces onboarding time and effort while eliminating errors. A “one-click” automated process accelerates modeling; automatically populates resource and environment parameters; and, offers a new extensible plug-in architecture.
Full-stack Dependency View – provides IT-Ops and App Support teams a way to model and visualize dependencies among applications, middleware and infrastructure. Users can define middleware and their infrastructure requirements per application; capture and report on infrastructure/middleware capabilities across environments; and, ensure dependencies are met.
Built-in IDE for CloudBees Flow DSL with Object Export enhancements__ – makes it easy to define, debug, version and store your “Process as Code” in any Source Control Management system. The built-in IDE provides an efficient way to modify application models, deployment processes and release pipelines as code and instantly see the changes reflected in the graphical interface.
We are excited about the new features in CloudBees Flow, and think you will be too. The new version of CloudBees Flow with Rolling Deployments is now available through our Invitation Program and will be generally available in August 2016. Learn more about CloudBees Flow, and be sure to download the Community edition to try the power of the platform, and use it for free on up to 10 hosts!