Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News GitLab 11.6 Supports Serverless Function Deployment

GitLab 11.6 Supports Serverless Function Deployment

Leia em Português

This item in japanese

The latest release of GitLab is able to use Knative and Kubernetes to build, deploy, and manage serverless workloads leveraging the Function as a service (FaaS) model.

With GitLab Serverless and support for FaaS, GitLab goes further down the road of extending its CI/CD capabilities into supporting a fuller DevOps lifecycle. GitLab introduced support for Knative in version 11.5 as an alpha feature, aiming to make it possible for developers to easily create serverless applications while reducing the risk of vendor lock-in. To this aim, GitLab leveraged the work done at TriggerMesh, which develops multi-cloud serverless and FaaS management solutions. GitLab 11.6 now adds support for deploying individual functions through a new UI for serverless operation which shows you a list of all your functions. The list includes a brief description and the Knative cluster it is deployed to.

(Image from GitLab blog)

Additionally, the GitLab Serverless UI can provide more detail for each deployed function, including the number of Kubernetes pods in use and the number of invocations over time.

(Image from GitLab blog)

According to GitLab,

By running serverless workloads on Kubernetes, businesses gain an abstraction layer that allows them to use compute resources from multiple cloud providers or even on-premises servers. This enhanced portability means they can choose the compute model that best meets their needs instead of being locked into a specific provider to run their functions.

FaaS enables the dynamical allocation of cloud resources to execute small units of code when an event occurs. Its main goal is auto-scaling your cloud infrastructure on-demand based on actual processing requests. FaaS removes the requirement of keeping at least one server process always running while you want to serve requests. Indeed, FaaS enables the server process to be spawned only when a request comes in. When there are not more requests to serve, the cluster scales down to zero. This makes it possible to lower cloud infrastructure costs, at the expense of setup latency. There are many FaaS providers, including Amazon, with AWS Lambda; Microsoft, offering Azure Functions; Google, with Cloud Functions, and others.

Other new features introduced in GitLab 11.6 are Suggested ChangesWeb Terminal for Web IDE, and the Group Security Dashboard Vulnerability Chart

Rate this Article