Add instance restart policy

cmd/api/api/instances.go

@@ -293,6 +293,13 @@ func (s *ApiService) CreateInstance(ctx context.Context, request oapi.CreateInst
 			Message: err.Error(),
 		}, nil
 	}
+	restartPolicy, err := toDomainRestartPolicy(request.Body.RestartPolicy)
+	if err != nil {
+		return oapi.CreateInstance400JSONResponse{
+			Code:    "invalid_restart_policy",
+			Message: err.Error(),
+		}, nil
+	}
 
 	domainReq := instances.CreateInstanceRequest{
 		Name:                     request.Body.Name,
@@ -319,6 +326,7 @@ func (s *ApiService) CreateInstance(ctx context.Context, request oapi.CreateInst
 		SkipGuestAgent:           request.Body.SkipGuestAgent != nil && *request.Body.SkipGuestAgent,
 		AutoStandby:              autoStandby,
 		HealthCheck:              healthCheck,
+		RestartPolicy:            restartPolicy,
 	}
 	if request.Body.SnapshotPolicy != nil {
 		snapshotPolicy, err := toInstanceSnapshotPolicy(*request.Body.SnapshotPolicy)
@@ -970,11 +978,20 @@ func (s *ApiService) UpdateInstance(ctx context.Context, request oapi.UpdateInst
 			Message: err.Error(),
 		}, nil
 	}
+	restartPolicy, err := toDomainRestartPolicy(request.Body.RestartPolicy)
+	if err != nil {
+		return oapi.UpdateInstance400JSONResponse{
+			Code:    "invalid_restart_policy",
+			Message: err.Error(),
+		}, nil
+	}
 
 	result, err := s.InstanceManager.UpdateInstance(ctx, inst.Id, instances.UpdateInstanceRequest{
-		Env:         env,
-		AutoStandby: autoStandby,
-		HealthCheck: healthCheck,
+		Env:              env,
+		AutoStandby:      autoStandby,
+		HealthCheck:      healthCheck,
+		RestartPolicy:    restartPolicy,
+		RestartPolicySet: request.Body.RestartPolicy != nil,
 	})
 	if err != nil {
 		switch {
@@ -1108,6 +1125,8 @@ func instanceToOAPI(inst instances.Instance) oapi.Instance {
 	oapiInst.AutoStandby = toOAPIAutoStandbyPolicy(inst.AutoStandby)
 	oapiInst.HealthCheck = toOAPIHealthCheck(inst.HealthCheck)
 	oapiInst.HealthStatus = toOAPIHealthStatus(healthcheck.Snapshot(inst.HealthCheck, string(inst.State), inst.HealthCheckRuntime))
+	oapiInst.RestartPolicy = toOAPIRestartPolicy(inst.RestartPolicy)
+	oapiInst.RestartStatus = toOAPIRestartStatus(inst.RestartStatus)
 
 	// Convert volume attachments
 	if len(inst.Volumes) > 0 {

cmd/api/api/instances_test.go

@@ -16,6 +16,7 @@ import (
 	mw "github.com/kernel/hypeman/lib/middleware"
 	"github.com/kernel/hypeman/lib/oapi"
 	"github.com/kernel/hypeman/lib/paths"
+	restartpolicy "github.com/kernel/hypeman/lib/restart-policy"
 	"github.com/kernel/hypeman/lib/system"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -281,6 +282,7 @@ func (m *captureUpdateManager) UpdateInstance(ctx context.Context, id string, re
 			Env:            req.Env,
 			AutoStandby:    req.AutoStandby,
 			HealthCheck:    req.HealthCheck,
+			RestartPolicy:  req.RestartPolicy,
 			CreatedAt:      now,
 			HypervisorType: hypervisor.TypeCloudHypervisor,
 		},
@@ -304,6 +306,7 @@ func (m *captureCreateManager) CreateInstance(ctx context.Context, req instances
 			Vcpus:          req.Vcpus,
 			AutoStandby:    req.AutoStandby,
 			HealthCheck:    req.HealthCheck,
+			RestartPolicy:  req.RestartPolicy,
 			CreatedAt:      now,
 			HypervisorType: hypervisor.TypeCloudHypervisor,
 		},
@@ -705,6 +708,48 @@ func TestCreateInstance_MapsHealthCheckPolicy(t *testing.T) {
 	assert.Equal(t, oapi.InstanceHealthStatusStatusStarting, instance.HealthStatus.Status)
 }
 
+func TestCreateInstance_MapsRestartPolicy(t *testing.T) {
+	t.Parallel()
+
+	svc := newTestService(t)
+	origMgr := svc.InstanceManager
+	mockMgr := &captureCreateManager{Manager: origMgr}
+	svc.InstanceManager = mockMgr
+
+	policy := oapi.OnFailure
+	backoff := "7s"
+	stableAfter := "2m"
+	maxAttempts := 4
+
+	resp, err := svc.CreateInstance(ctx(), oapi.CreateInstanceRequestObject{
+		Body: &oapi.CreateInstanceRequest{
+			Name:  "test-restart-policy",
+			Image: "docker.io/library/alpine:latest",
+			RestartPolicy: &oapi.RestartPolicy{
+				Policy:      &policy,
+				Backoff:     &backoff,
+				StableAfter: &stableAfter,
+				MaxAttempts: &maxAttempts,
+			},
+		},
+	})
+	require.NoError(t, err)
+
+	created, ok := resp.(oapi.CreateInstance201JSONResponse)
+	require.True(t, ok, "expected 201 response")
+	require.NotNil(t, mockMgr.lastReq)
+	require.NotNil(t, mockMgr.lastReq.RestartPolicy)
+	assert.Equal(t, restartpolicy.PolicyOnFailure, mockMgr.lastReq.RestartPolicy.Policy)
+	assert.Equal(t, "7s", mockMgr.lastReq.RestartPolicy.Backoff)
+	assert.Equal(t, "2m0s", mockMgr.lastReq.RestartPolicy.StableAfter)
+	assert.Equal(t, 4, mockMgr.lastReq.RestartPolicy.MaxAttempts)
+
+	instance := oapi.Instance(created)
+	require.NotNil(t, instance.RestartPolicy)
+	require.NotNil(t, instance.RestartPolicy.Policy)
+	assert.Equal(t, oapi.OnFailure, *instance.RestartPolicy.Policy)
+}
+
 func TestUpdateInstance_MapsEnvPatch(t *testing.T) {
 	t.Parallel()
 	svc := newTestService(t)
@@ -883,6 +928,108 @@ func TestUpdateInstance_MapsHealthCheckPatch(t *testing.T) {
 	assert.Equal(t, oapi.InstanceHealthStatusStatusUnknown, instance.HealthStatus.Status)
 }
 
+func TestUpdateInstance_MapsRestartPolicyPatch(t *testing.T) {
+	t.Parallel()
+	svc := newTestService(t)
+
+	origMgr := svc.InstanceManager
+	now := time.Now()
+	mockMgr := &captureUpdateManager{
+		Manager: origMgr,
+		result: &instances.Instance{
+			StoredMetadata: instances.StoredMetadata{
+				Id:             "inst-update-restart-policy",
+				Name:           "inst-update-restart-policy",
+				Image:          "docker.io/library/alpine:latest",
+				CreatedAt:      now,
+				HypervisorType: hypervisor.TypeCloudHypervisor,
+				RestartPolicy: &restartpolicy.Policy{
+					Policy:      restartpolicy.PolicyAlways,
+					Backoff:     "5s",
+					StableAfter: "10m0s",
+				},
+				RestartStatus: restartpolicy.Status{
+					BlockedReason: restartpolicy.BlockedReasonManualStop,
+				},
+			},
+			State: instances.StateStopped,
+		},
+	}
+	svc.InstanceManager = mockMgr
+
+	policy := oapi.Always
+	resolved := &instances.Instance{
+		StoredMetadata: instances.StoredMetadata{
+			Id:             "inst-update-restart-policy",
+			Name:           "inst-update-restart-policy",
+			Image:          "docker.io/library/alpine:latest",
+			CreatedAt:      now,
+			HypervisorType: hypervisor.TypeCloudHypervisor,
+		},
+		State: instances.StateStopped,
+	}
+
+	resp, err := svc.UpdateInstance(mw.WithResolvedInstance(ctx(), resolved.Id, resolved), oapi.UpdateInstanceRequestObject{
+		Id: resolved.Id,
+		Body: &oapi.UpdateInstanceRequest{
+			RestartPolicy: &oapi.RestartPolicy{Policy: &policy},
+		},
+	})
+	require.NoError(t, err)
+	updated, ok := resp.(oapi.UpdateInstance200JSONResponse)
+	require.True(t, ok, "expected 200 response")
+
+	require.NotNil(t, mockMgr.lastReq)
+	assert.True(t, mockMgr.lastReq.RestartPolicySet)
+	require.NotNil(t, mockMgr.lastReq.RestartPolicy)
+	assert.Equal(t, restartpolicy.PolicyAlways, mockMgr.lastReq.RestartPolicy.Policy)
+
+	instance := oapi.Instance(updated)
+	require.NotNil(t, instance.RestartPolicy)
+	require.NotNil(t, instance.RestartStatus)
+	require.NotNil(t, instance.RestartStatus.BlockedReason)
+	assert.Equal(t, oapi.ManualStop, *instance.RestartStatus.BlockedReason)
+}
+
+func TestUpdateInstance_RejectsInvalidRestartPolicy(t *testing.T) {
+	t.Parallel()
+	svc := newTestService(t)
+
+	origMgr := svc.InstanceManager
+	mockMgr := &captureUpdateManager{Manager: origMgr}
+	svc.InstanceManager = mockMgr
+
+	now := time.Now()
+	resolved := &instances.Instance{
+		StoredMetadata: instances.StoredMetadata{
+			Id:             "inst-update-restart-policy",
+			Name:           "inst-update-restart-policy",
+			Image:          "docker.io/library/alpine:latest",
+			CreatedAt:      now,
+			HypervisorType: hypervisor.TypeCloudHypervisor,
+		},
+		State: instances.StateStopped,
+	}
+	policy := oapi.OnFailure
+	backoff := "0s"
+
+	resp, err := svc.UpdateInstance(mw.WithResolvedInstance(ctx(), resolved.Id, resolved), oapi.UpdateInstanceRequestObject{
+		Id: resolved.Id,
+		Body: &oapi.UpdateInstanceRequest{
+			RestartPolicy: &oapi.RestartPolicy{
+				Policy:  &policy,
+				Backoff: &backoff,
+			},
+		},
+	})
+	require.NoError(t, err)
+
+	badReq, ok := resp.(oapi.UpdateInstance400JSONResponse)
+	require.True(t, ok, "expected 400 response")
+	assert.Equal(t, "invalid_restart_policy", badReq.Code)
+	assert.Nil(t, mockMgr.lastReq)
+}
+
 func TestUpdateInstance_RejectsZeroAutoStandbyIgnoreDestinationPort(t *testing.T) {
 	t.Parallel()
 	svc := newTestService(t)

cmd/api/api/restart_policy.go

@@ -0,0 +1,80 @@
+package api
+
+import (
+	"github.com/kernel/hypeman/lib/oapi"
+	restartpolicy "github.com/kernel/hypeman/lib/restart-policy"
+	"github.com/samber/lo"
+)
+
+func toDomainRestartPolicy(policy *oapi.RestartPolicy) (*restartpolicy.Policy, error) {
+	if policy == nil {
+		return nil, nil
+	}
+
+	out := &restartpolicy.Policy{}
+	if policy.Policy != nil {
+		out.Policy = restartpolicy.PolicyMode(*policy.Policy)
+	}
+	if policy.Backoff != nil {
+		out.Backoff = *policy.Backoff
+	}
+	if policy.MaxAttempts != nil {
+		out.MaxAttempts = *policy.MaxAttempts
+	}
+	if policy.StableAfter != nil {
+		out.StableAfter = *policy.StableAfter
+	}
+	normalized, err := restartpolicy.NormalizePolicy(out)
+	if err != nil {
+		return nil, err
+	}
+	return normalized, nil
+}
+
+func toOAPIRestartPolicy(policy *restartpolicy.Policy) *oapi.RestartPolicy {
+	if policy == nil {
+		return nil
+	}
+
+	mode := oapi.RestartPolicyPolicy(policy.Policy)
+	out := &oapi.RestartPolicy{
+		Policy: &mode,
+	}
+	if policy.Backoff != "" {
+		out.Backoff = lo.ToPtr(policy.Backoff)
+	}
+	if policy.MaxAttempts > 0 {
+		out.MaxAttempts = lo.ToPtr(policy.MaxAttempts)
+	}
+	if policy.StableAfter != "" {
+		out.StableAfter = lo.ToPtr(policy.StableAfter)
+	}
+	return out
+}
+
+func toOAPIRestartStatus(status restartpolicy.Status) *oapi.RestartStatus {
+	if status.IsZero() {
+		return nil
+	}
+
+	out := &oapi.RestartStatus{
+		Attempts: lo.ToPtr(status.Attempts),
+	}
+	if status.BlockedReason != "" {
+		reason := oapi.RestartStatusBlockedReason(status.BlockedReason)
+		out.BlockedReason = &reason
+	}
+	if status.LastAttemptAt != nil {
+		lastAttemptAt := status.LastAttemptAt.UTC()
+		out.LastAttemptAt = &lastAttemptAt
+	}
+	if status.NextAttemptAt != nil {
+		nextAttemptAt := status.NextAttemptAt.UTC()
+		out.NextAttemptAt = &nextAttemptAt
+	}
+	if status.LastReason != "" {
+		reason := oapi.RestartStatusLastReason(status.LastReason)
+		out.LastReason = &reason
+	}
+	return out
+}

cmd/api/main.go

@@ -571,6 +571,14 @@ func run() error {
 			return app.HealthCheckController.Run(gctx)
 		})
 	}
+	if restartController, ok := app.InstanceManager.(interface {
+		StartRestartPolicyController(context.Context) error
+	}); ok {
+		grp.Go(func() error {
+			logger.Info("starting restart policy controller")
+			return restartController.StartRestartPolicyController(gctx)
+		})
+	}
 
 	// Run the server
 	grp.Go(func() error {

lib/healthcheck/README.md

@@ -75,6 +75,10 @@ Stopping, deleting, standing by, or restoring an instance stops active checks. S
 
 ## Restart Policy
 
-Health checks only report health. They do not restart instances.
+Health checks do not restart instances by themselves.
 
-If Hypeman later adds restart-on-unhealthy behavior, it should consume `health_status=unhealthy` explicitly rather than making health checks mutate lifecycle state.
+When an instance also has `restart_policy.policy=on_failure` or `restart_policy.policy=always`, an `unhealthy` health status becomes a restart-policy failure signal. The restart policy applies its normal backoff, max attempts, manual-stop suppression, and stable-window reset before Hypeman restarts the whole instance.
+
+With `restart_policy.policy=never` or no restart policy, health checks only report status.
+
+Health checks still do not mutate lifecycle state directly. The instance remains `Running` while unhealthy until restart policy chooses to stop and start it.

lib/instances/create.go

@@ -375,6 +375,7 @@ func (m *manager) createInstance(
 		SnapshotPolicy:           cloneSnapshotPolicy(req.SnapshotPolicy),
 		AutoStandby:              cloneAutoStandbyPolicy(req.AutoStandby),
 		HealthCheck:              cloneHealthCheckPolicy(req.HealthCheck),
+		RestartPolicy:            cloneRestartPolicy(req.RestartPolicy),
 	}
 
 	// 12. Ensure directories
@@ -638,6 +639,11 @@ func validateCreateRequest(req *CreateInstanceRequest) error {
 	if err := validateHealthCheckCompatibility(req.HealthCheck, req.NetworkEnabled, req.SkipGuestAgent); err != nil {
 		return err
 	}
+	normalizedRestartPolicy, err := normalizeRestartPolicy(req.RestartPolicy)
+	if err != nil {
+		return err
+	}
+	req.RestartPolicy = normalizedRestartPolicy
 
 	// Validate volume attachments
 	if err := validateVolumeAttachments(req.Volumes); err != nil {

lib/instances/fork.go

@@ -16,6 +16,7 @@ import (
 	"github.com/kernel/hypeman/lib/instances/phasetracking"
 	"github.com/kernel/hypeman/lib/logger"
 	"github.com/kernel/hypeman/lib/network"
+	restartpolicy "github.com/kernel/hypeman/lib/restart-policy"
 	"github.com/nrednav/cuid2"
 	"go.opentelemetry.io/otel/attribute"
 	"gvisor.dev/gvisor/pkg/cleanup"
@@ -281,6 +282,7 @@ func (m *manager) forkInstanceFromStoppedOrStandby(ctx context.Context, id strin
 	forkMeta.VsockSocket = m.paths.InstanceSocket(forkID, hypervisor.VsockSocketNameForType(forkMeta.HypervisorType))
 	forkMeta.ExitCode = nil
 	forkMeta.ExitMessage = ""
+	forkMeta.RestartStatus = restartpolicy.Status{}
 	// Forks are new instances; phase accounting must not inherit the source's
 	// cumulative durations. The first transition into the fork's runtime
 	// phase (Standby for snapshot forks, Stopped for stopped forks) will be
@@ -481,6 +483,9 @@ func cloneStoredMetadata(src StoredMetadata) StoredMetadata {
 	if src.HealthCheck != nil {
 		dst.HealthCheck = cloneHealthCheckPolicy(src.HealthCheck)
 	}
+	if src.RestartPolicy != nil {
+		dst.RestartPolicy = cloneRestartPolicy(src.RestartPolicy)
+	}
 	if src.SnapshotPolicy != nil {
 		dst.SnapshotPolicy = cloneSnapshotPolicy(src.SnapshotPolicy)
 	}