Alloc Health should take Group Checks into Consideration

[schmichael]

This code comment is incorrect:

nomad/client/allochealth/tracker.go

Lines 241 to 246 in 0ecda99

	//TODO(schmichael) for now skip unknown tasks as
	//they're task group services which don't currently
	//support checks anyway
	if v, ok := t.taskHealth[task]; ok {
	v.state = state
	}

Group checks are supported and should be taken into consideration when setting allocation health.

[notnoop]

I have tested the code and it looks we properly account for service check health.

The group checks get presented as associated with a "fake" task with the group name as the task name.

The relevant bits are:
[1]

nomad/client/allocrunner/groupservice_hook.go

Line 94 in 2949c44

return h.consulClient.RegisterWorkload(services)

[2]

nomad/client/allocrunner/groupservice_hook.go

Lines 192 to 206 in 2949c44

	func (h *groupServiceHook) getWorkloadServices() *agentconsul.WorkloadServices {
	// Interpolate with the task's environment
	interpolatedServices := taskenv.InterpolateServices(h.taskEnvBuilder.Build(), h.services)
	// Create task services struct with request's driver metadata
	return &agentconsul.WorkloadServices{
	AllocID: h.allocID,
	Group: h.group,
	Restarter: h.restarter,
	Services: interpolatedServices,
	DriverNetwork: h.driverNet(),
	Networks: h.networks,
	Canary: h.canary,
	}
	}

[3]

nomad/command/agent/consul/client.go

Line 942 in 2949c44

c.addRegistrations(workload.AllocID, workload.Name(), t)

[4]

nomad/command/agent/consul/structs.go

Lines 88 to 94 in 2949c44

	func (w *WorkloadServices) Name() string {
	if w.Task != "" {
	return w.Task
	}
	return "group-" + w.Group
	}

Not sure I fully understand the TODO in code referenced in the issue. Group services don't have task health associated with them. But should they? Should we track the envoy/sidecar containers as task health? Task health differ from checks as they simply check if the task is running.

[notnoop]

Also, we probably should document that consul registrations may contain synthetic tasks not present in job description - it'd be an easy mistake to make - like I've done in https://github.com/hashicorp/nomad/pull/7366/files .

[github-actions]

I'm going to lock this issue because it has been closed for 120 days ⏳. This helps our maintainers find and focus on the active issues.
If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.