Chapter 4. Adding Workloads to the Mesh
Getting Linkerd running in your cluster is a great first step. But it’s pointless to run Linkerd with nothing else: to get actual value out of your Linkerd cluster, you’ll need to get workloads running in your service mesh. In this chapter, we’ll show you how to do exactly that.
Workloads Versus Services
We’ll talk about “workloads” a lot in this chapter—but sometimes we’ll also talk about “services,” and sometimes “Services.” Unfortunately, these three things all have slightly different meanings:
- Service
-
A Kubernetes resource that is widely used to control how Kubernetes allocates DNS names and IP addresses for services (see Figure 4-1).
- Workload
-
A thing that actually does work on your behalf. A workload receives requests over the network and executes code to take actions. In Kubernetes, it’s usually one or more Pods (to provide the computation), often managed by a Deployment or DaemonSet resource, plus one or more Services (to manage the names and IP addresses), as illustrated in Figure 4-1.
- service
-
A less formal term that could refer to either a Service or a workload, depending on context. This lack of precision shows just one of many cases where Kubernetes terminology is very much more confusing than we would like it to be.
Figure 4-1. The workload as distinct from the Service
As an application developer, you can usually just ...