Frequently Asked Questions

What is Linkerd?

Linkerd is a servicemesh.It adds observability, reliability, and security to cloud native applications,without requiring code changes. For example, Linkerd can monitor and reportper-service success rates and latencies, can automatically retry failedrequests, and can encrypt and validate connections between services, allwithout requiring any modification of the application itself.

Linkerd works by inserting ultralight proxies (collectively, the “data plane”)alongside each application instance. Linkerd’s control plane provides operatorswith a uniform point at which they can control and measure the behavior of thedata plane. Operators typically interact with Linkerd using the CLIand the web dashboard UI.

Who owns Linkerd and how is it licensed?

Linkerd is hosted by the Cloud Native Computing Foundation(CNCF) project. The CNCF owns the trademark; the copyright is held by theLinkerd authors themselves.

Linkerd is licensed under Apache2.0.

Who is Linkerd for?

Linkerd is for everyone. In practice, Linkerd has certain technicalprerequisites. Read on below.

How do I pronounce Linkerd?

The “d” is pronounced separately, i.e. “Linker-DEE”. (It’s a UNIX thing.)

How do I write Linkerd?

Just like this: Linkerd. Capital “L”, lower-case everything else.

Who maintains Linkerd?

See the 2.xmaintainersfile, and the 1.xmaintainersfile.

Is there an Enterprise edition, or a commercial edition?

No. Everything in Linkerd is fully open source. Some companies providecommercial support for Linkerd.

What’s the difference between Linkerd 1.x and 2.x?

Linkerd 1.x is built on the “Twitter stack”: Finagle, Netty, Scala, and theJVM.

Linkerd 2.x is built in Rust and Go. It is significantly faster andlighter weight than 1.x, but currently only supports Kubernetes.

Is Linkerd 1.x still supported?

Yes, the 1.x branch of Linkerd is under active development, and continuesto power the production infrastructure of companies around the globe.

The full Linkerd 1.x documentation is here.

Does Linkerd require Kubernetes?

Linkerd 2.x currently requires Kubernetes, though this will change in thefuture. Linkerd 1.x can be installed on any platform, and supports Kubernetes,DC/OS, Mesos, Consul, and ZooKeeper-based environments.

Where’s the Linkerd roadmap?

As a community project, there is no official roadmap. A glance at the activeGitHub issues will give you asense of what is in store for the future.

The public Linkerd Meetupslidesalso provides a coarse-grained roadmap.

How secure is Linkerd?

Very secure. A third party security audit of Linkerd2.x wascompleted in June 2019, and Linkerd passed with flying colors.

How fast is Linkerd?

Very fast. A third party performance evaluation of Linkerd vsIstio was performed in Mayof 2019, and showed that Linkerd significantly outperformed Istio.

How do I use Linkerd to route traffic between services?

(2.x) Linkerd doesn’t change the way routing works. You address other services justas you would without Linkerd, e.g. service-name.namespace.svc.cluster.local,or service-name if within the same namespace.

How does Linkerd handle ingress?

(2.x) For reasons of simplicity, Linkerd doesn’t provide ingress itself, butinstead works in conjunction with the ingresscontroller of your choice.

What happens to Linkerd’s proxies if the control plane is down?

Linkerd’s proxies do not integrate with Kubernetes directly, but rely on thecontrol plane for service discovery information. The proxies are designed tocontinue operating even if they can’t reach the control plane.

If the control plane dies, existing proxies will continue to operate with thelatest service discovery information. If Additionally, they will fall back toDNS if asked to route to a service they don’t have information for. (Thus, ifthe control plane is down, but new services are created, you may noticedifferent load balancing behavior until the control plane resumes.) Once thecontrol plane is functional, the Linkerd proxies will resume communication asnormal.

If new proxies are deployed when the control plane is unreachable, these newproxies will not be able to operate. They will timeout all new requests untilsuch time as they can reach the control plane.