feat(notify): upgrade dingtalk alert notifications to actioncard and markdown with rich chinese formatting

cmd/ongrid/main.go

@@ -2940,6 +2940,7 @@ func main() {
 			PromQuerier:     alertPromQuerier,
 			LogQuerier:      lokiLogClient,
 			LogSearcher:     logsBackendSvc,
+			ConsoleURL:      cfg.PublicURL,
 			DeviceIdentityResolver: func(ctx context.Context, deviceID uint64) (managerbizalert.DeviceIdentity, error) {
 				device, err := deviceUC.Get(ctx, deviceID)
 				if err != nil {
@@ -2964,6 +2965,7 @@ func main() {
 			Usecase:     alertUC,
 			MaxAttempts: 5,
 			Tick:        cfg.Alert.EvaluatorInterval,
+			ConsoleURL:  cfg.PublicURL,
 			Log:         log.With(slog.String("comp", "alert-retry")),
 		})
 		eg.Go(func() error { return retryWorker.Loop(egCtx) })

internal/manager/biz/alert/pipeline.go

@@ -91,6 +91,10 @@ type PipelineEvaluatorOpts struct {
 
 	DeviceIdentityResolver DeviceIdentityResolver
 
+	// ConsoleURL is the public console base used for notification card deep links
+	// (devices / incidents). Empty disables action buttons.
+	ConsoleURL string
+
 	Log *slog.Logger
 	Now func() time.Time
 }
@@ -117,6 +121,7 @@ type PipelineEvaluator struct {
 	logq           LogQuerier
 	logSearcher    logquery.Searcher
 	deviceIdentity DeviceIdentityResolver
+	consoleURL     string
 
 	// gaugeSnapshot is the previous tick's (device_id, device_name) set
 	// used by refreshDeviceStalenessGauge to garbage-collect series for
@@ -166,6 +171,7 @@ func NewPipelineEvaluator(opts PipelineEvaluatorOpts) *PipelineEvaluator {
 		logq:           opts.LogQuerier,
 		logSearcher:    opts.LogSearcher,
 		deviceIdentity: opts.DeviceIdentityResolver,
+		consoleURL:     strings.TrimRight(strings.TrimSpace(opts.ConsoleURL), "/"),
 		log:            opts.Log,
 		now:            opts.Now,
 	}
@@ -423,16 +429,24 @@ func (e *PipelineEvaluator) notify(ctx context.Context, res *FiringResult, summa
 		OccurredAt: at,
 		Labels: map[string]string{
 			"rule":        res.Incident.Rule,
+			"rule_name":   res.Incident.RuleName,
+			"status":      res.Incident.Status,
 			"incident_id": fmt.Sprintf("%d", res.Incident.ID),
 		},
 	}
+	if res.Incident.RunbookURL != "" {
+		msg.Labels["runbook_url"] = res.Incident.RunbookURL
+	}
+	if e.consoleURL != "" {
+		msg.Labels["console_url"] = e.consoleURL
+	}
 	// Preserve application identity for notification routing and correlation.
 	// Control labels such as incident_id/rule still come from the incident.
 	if labels, err := res.Incident.Labels(); err != nil {
 		e.log.Warn("alert: decode notification identity failed", slog.Uint64("incident_id", res.Incident.ID), slog.Any("err", err))
 	} else {
-		for _, key := range []string{"service", "service_namespace", "deployment_environment_name", "service_instance_id", "span_name"} {
-			if value, ok := labels[key]; ok {
+		for _, key := range []string{"service", "service_namespace", "deployment_environment_name", "service_instance_id", "span_name", "rule_name", "rule_expr"} {
+			if value, ok := labels[key]; ok && msg.Labels[key] == "" {
 				msg.Labels[key] = value
 			}
 		}
@@ -446,6 +460,7 @@ func (e *PipelineEvaluator) notify(ctx context.Context, res *FiringResult, summa
 				e.log.Warn("alert: resolve device identity for notification failed",
 					slog.Uint64("device_id", deviceID), slog.Any("err", err))
 			} else if display := deviceDisplay(identity); display != "" {
+				msg.Labels["device_name"] = display
 				msg.Subject = strings.ReplaceAll(msg.Subject,
 					fmt.Sprintf("device_id=%d", deviceID), "device="+display)
 				if identity.Hostname != "" {

internal/manager/biz/alert/pipeline_test.go

@@ -39,6 +39,30 @@ func TestPipelineNotificationPreservesApplicationIdentity(t *testing.T) {
 	}
 }
 
+func TestPipelineNotificationInjectsConsoleURL(t *testing.T) {
+	repo, notifier := newFakeRepo(), &fakeNotifier{}
+	prom := &fakePromQuerier{result: &promquery.InstantResult{ResultType: "vector", Result: json.RawMessage(`[{"metric":{"device_id":"9"},"value":[1,"1"]}]`)}}
+	rules := NewStaticRulesProvider(WithMetricRawRules([]MetricRawRule{{ID: 1, RuleKey: "cpu_high", Name: "CPU high", ScopeType: "global", Expr: "up > 0", Severity: "critical"}}))
+	evaluator := newPipelineEvaluator(t, repo, notifier, rules, PipelineEvaluatorOpts{
+		PromQuerier: prom,
+		ConsoleURL:  "https://ongrid.example.com/",
+	})
+	evaluator.EvaluateOnce(t.Context())
+	if len(notifier.msgs) != 1 {
+		t.Fatalf("notifications=%d", len(notifier.msgs))
+	}
+	labels := notifier.msgs[0].Labels
+	if labels["console_url"] != "https://ongrid.example.com" {
+		t.Fatalf("console_url = %q, want trimmed public URL", labels["console_url"])
+	}
+	if labels["status"] != model.IncidentStatusOpen {
+		t.Fatalf("status = %q, want %q", labels["status"], model.IncidentStatusOpen)
+	}
+	if labels["rule_name"] != "CPU high" {
+		t.Fatalf("rule_name = %q", labels["rule_name"])
+	}
+}
+
 func (f *fakeEdgeLister) List(_ context.Context, _ edgebiz.ListFilter) ([]*edgemodel.Edge, error) {
 	return f.edges, f.err
 }

internal/manager/biz/alert/retry.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"fmt"
 	"log/slog"
+	"strings"
 	"time"
 
 	model "github.com/ongridio/ongrid/internal/manager/model/alert"
@@ -23,8 +24,11 @@ type RetryWorkerOpts struct {
 	MaxAttempts       uint32
 	BackoffPerAttempt time.Duration
 	Tick              time.Duration
-	Log               *slog.Logger
-	Now               func() time.Time
+	// ConsoleURL is the public console base injected into retry payloads so
+	// Feishu cards retain the same deep-link buttons as first-send notifies.
+	ConsoleURL string
+	Log        *slog.Logger
+	Now        func() time.Time
 }
 
 // RetryWorker drains failed notification_deliveries, re-runs the underlying
@@ -38,6 +42,7 @@ type RetryWorker struct {
 	maxAttempts       uint32
 	backoffPerAttempt time.Duration
 	tick              time.Duration
+	consoleURL        string
 	log               *slog.Logger
 	now               func() time.Time
 }
@@ -67,6 +72,7 @@ func NewRetryWorker(opts RetryWorkerOpts) *RetryWorker {
 		maxAttempts:       opts.MaxAttempts,
 		backoffPerAttempt: opts.BackoffPerAttempt,
 		tick:              opts.Tick,
+		consoleURL:        strings.TrimRight(strings.TrimSpace(opts.ConsoleURL), "/"),
 		log:               opts.Log,
 		now:               opts.Now,
 	}
@@ -158,7 +164,7 @@ func (w *RetryWorker) retryOne(ctx context.Context, d *model.Delivery, now time.
 		return
 	}
 
-	msg := buildIncidentMessage(incident, now)
+	msg := buildIncidentMessage(incident, now, w.consoleURL)
 	sentAt := now
 	finished := now
 	sendErr := w.notifier.Send(ctx, msg, channel.Name)
@@ -200,15 +206,25 @@ func (w *RetryWorker) retryOne(ctx context.Context, d *model.Delivery, now time.
 	}
 }
 
-func buildIncidentMessage(incident *model.Incident, now time.Time) notify.Message {
+func buildIncidentMessage(incident *model.Incident, now time.Time, consoleURL string) notify.Message {
 	severity := notify.Severity(incident.Severity)
 	if severity == "" {
 		severity = notify.SeverityWarning
 	}
 	labels := map[string]string{
 		"rule":        incident.Rule,
+		"status":      incident.Status,
 		"incident_id": fmt.Sprintf("%d", incident.ID),
 	}
+	if incident.RuleName != "" {
+		labels["rule_name"] = incident.RuleName
+	}
+	if incident.RunbookURL != "" {
+		labels["runbook_url"] = incident.RunbookURL
+	}
+	if consoleURL != "" {
+		labels["console_url"] = consoleURL
+	}
 	if incident.DeviceID != nil {
 		labels["device_id"] = fmt.Sprintf("%d", *incident.DeviceID)
 	}