Mock Fleet Server

It's a mock for fleet-server, allowing testing of Agent interactions with fleet-server without the need to run a fleet-server and have full control over it, to even test edge cases such as error handling.

The server is designed to work with one single agent. The Agent either needs to be enrolled or AgentID must be set on Handlers. Use WithAgentID to set it when creating the server. Also, the requests are executed sequentially, making safe for the handlers implementations to access and eventually change the Handlers properties.

Right now there is no authentication, which means any enrollment token will work by default. There is some work done for authentication on auth.go.

tl;dr

• See testing/integration/proxy_url_test.go

• and fleetservertest_test.go for examples.

• on fleetservertest.Handlers the handlernameFn properties are used for

• implementing the handlers.By default, any not implemented handler returns a

• http.StatusNotImplemented.

• Use fleetservertest.NewServer(fleetservertest.Handlers{}) to create a new

• test server. It's a *httptest.Server:

	NewServer(&Handlers{
		AckFn:            nil,
		CheckinFn:        nil,
		EnrollFn:         nil,
		ArtifactFn:       nil,
		StatusFn:         nil,
		UploadBeginFn:    nil,
		UploadChunkFn:    nil,
		UploadCompleteFn: nil,
	})

• Use the fleetservertest.NewPATHNAME(args) functions to get a path ready to be used:

p := NewPathAgentAcks("my-agent-id")
// p = "/api/fleet/agents/my-agent-id/acks"

• Use fleetservertest.NewHANDERNAME() to get a ready to use handler:

ts := fleetservertest.NewServer(&Handlers{
	CheckinFn: fleetservertest.NewHandlerStatusHealth(),
})

• Check proxy_url_test.go for examples.
• Check fleetserver_test.go for examples.
• Check handlers.go for the available paths and handlers.
• Check models.go for the request and response models or the openapi definition.