Keep what each machine says of what it runs, raise it, and gate on it (hq ADR 0240, to-be 48 Phase A)

The gate judged a module by what the mesh saw from outside, so a container that
crash-looped after it applied passed it. Each machine's node-engine now states
the health of every long-running resource it runs; the controller keeps the
newest statement per machine, raises module.<module>.<machine>.unhealthy on the
second statement in a row, clears it on the first that does not say it, and the
gate passes a module only when every long-running resource of it is stated
healthy since the send. An engine that states nothing is judged as before.
This commit is contained in:
jochen
2026-10-07 02:28:16 +02:00
parent 5d3e52219b
commit 1cc6a2d759
22 changed files with 997 additions and 5 deletions
+42
View File
@@ -346,6 +346,48 @@ func TestTheHealthDefinitions(t *testing.T) {
!strings.Contains(why, "first run") {
t.Errorf("a controller not ready is %v (%s)", h, why)
}
// What the module runs (novox/hq ADR 0240 §4): a judging passes only when every long-running
// resource of it on that machine is stated healthy since the send; starting and unhealthy wait.
f = applied("anchor")
stated := func(state, reason string, heard time.Time) {
f.health = map[string]inventory.NodeHealth{"anchor": {Node: "anchor", Contract: 1, SaidAt: heard, HeardAt: heard,
Resources: []inventory.ResourceHealth{
{Module: "app", Resource: "app.server", Kind: "container", Target: "app-server", State: state, Reason: reason},
{Module: "other", Resource: "other.server", Kind: "container", State: link.StateUnhealthy, Reason: "down"}}}}
}
plain := catalogue.Manifest{Module: "app"}
if h, why := judgeHealth("app", "", plain, "anchor", since, f); h != healthGood {
t.Errorf("a machine whose engine states no health is judged as before, not %v (%s)", h, why)
}
stated(link.StateStarting, "", now)
if h, why := judgeHealth("app", "", plain, "anchor", since, f); h != healthNotYet || !strings.Contains(why, "starting") {
t.Errorf("a resource still starting is not yet a pass: %v (%s)", h, why)
}
stated(link.StateUnhealthy, "restarting", now)
if h, why := judgeHealth("app", "", plain, "anchor", since, f); h != healthNotYet || !strings.Contains(why, "restarting") {
t.Errorf("a resource unhealthy fails the judging: %v (%s)", h, why)
}
stated(link.StateHealthy, "", since.Add(-time.Minute))
if h, why := judgeHealth("app", "", plain, "anchor", since, f); h != healthNotYet {
t.Errorf("a statement from before the send says nothing of the new build: %v (%s)", h, why)
}
stated(link.StateHealthy, "", now)
if h, why := judgeHealth("app", "", plain, "anchor", since, f); h != healthGood {
t.Errorf("every resource of it healthy since the send — another module's state is not its — is %v (%s)", h, why)
}
f.healthErr = errors.New("the store is away")
if h, _ := judgeHealth("app", "", plain, "anchor", since, f); h != healthNotYet {
t.Errorf("health that cannot be read is never read as healthy: %v", h)
}
// And the condition it raises after the send holds it, as every condition about it does.
f.healthErr = nil
f.judged = true
f.open = []conditions.Condition{{Key: "module.app.anchor.unhealthy", Kind: kindModuleUnhealthy, Subject: conditions.Subject{
Scope: conditions.ScopeModule, ID: "app.anchor", Machine: "anchor"}, Raised: now, Summary: "app on anchor is not healthy"}}
if h, why := judgeHealth("app", "", plain, "anchor", since, f); h != healthNotYet || !strings.Contains(why, "module.app.anchor.unhealthy") {
t.Errorf("a module held on its own unhealthy condition is %v (%s)", h, why)
}
// A machine that refused what it was sent: broken.
f = applied("anchor")
r := f.reports["anchor"]