Add entra ID metrics

lib/msgraph/client.go

@@ -37,6 +37,7 @@ import (
 	"github.com/google/uuid"
 	"github.com/gravitational/trace"
 	"github.com/jonboulle/clockwork"
+	"github.com/prometheus/client_golang/prometheus"
 
 	"github.com/gravitational/teleport"
 	apidefaults "github.com/gravitational/teleport/api/defaults"
@@ -98,6 +99,9 @@ type Config struct {
 	// GraphEndpoint specifies root domain of the Graph API.
 	GraphEndpoint string
 	Logger        *slog.Logger
+	// MetricsRegistry configures where metrics should be registered.
+	// When nil, metrics are created but not registered.
+	MetricsRegistry prometheus.Registerer
 }
 
 // SetDefaults sets the default values for optional fields.
@@ -144,6 +148,7 @@ type Client struct {
 	baseURL       *url.URL
 	pageSize      int
 	logger        *slog.Logger
+	metrics       *clientMetrics
 }
 
 // NewClient returns a new client for the given config.
@@ -156,6 +161,15 @@ func NewClient(cfg Config) (*Client, error) {
 	if err != nil {
 		return nil, trace.Wrap(err)
 	}
+
+	metrics := newMetrics()
+	// gracefully handle not being given a metric registry
+	if cfg.MetricsRegistry != nil {
+		if err := metrics.register(cfg.MetricsRegistry); err != nil {
+			return nil, trace.Wrap(err, "registering metrics")
+		}
+	}
+
 	return &Client{
 		httpClient:    cfg.HTTPClient,
 		tokenProvider: cfg.TokenProvider,
@@ -164,6 +178,7 @@ func NewClient(cfg Config) (*Client, error) {
 		baseURL:       base.JoinPath(graphVersion),
 		pageSize:      cfg.PageSize,
 		logger:        cfg.Logger,
+		metrics:       metrics,
 	}, nil
 }
 
@@ -201,6 +216,7 @@ func (c *Client) request(ctx context.Context, method string, uri string, header
 	}
 
 	var lastErr error
+	var start time.Time
 	for range maxRetries {
 		if retryAfter > 0 {
 			select {
@@ -231,10 +247,13 @@ func (c *Client) request(ctx context.Context, method string, uri string, header
 		// https://learn.microsoft.com/en-us/graph/best-practices-concept#reliability-and-support
 		req.Header.Set("client-request-id", requestID)
 
+		start = c.clock.Now()
 		resp, err := c.httpClient.Do(req)
 		if err != nil {
 			return nil, trace.Wrap(err) // hard I/O error, bail
 		}
+		c.metrics.requestDuration.WithLabelValues(method).Observe(c.clock.Since(start).Seconds())
+		c.metrics.requestTotal.WithLabelValues(method, strconv.Itoa(resp.StatusCode))
 
 		if resp.StatusCode >= 200 && resp.StatusCode < 400 {
 			return resp, nil

lib/msgraph/metrics.go

@@ -0,0 +1,64 @@
+/*
+ * Teleport
+ * Copyright (C) 2025  Gravitational, Inc.
+ *
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+package msgraph
+
+import (
+	"github.com/gravitational/trace"
+	"github.com/prometheus/client_golang/prometheus"
+
+	"github.com/gravitational/teleport"
+)
+
+type clientMetrics struct {
+	// requestsTotal keeps track of the number of requests done by the client
+	// This metric is labeled by status code.
+	requestTotal *prometheus.CounterVec
+	// requestDuration keeps track of the request duration, in seconds.
+	requestDuration *prometheus.HistogramVec
+}
+
+const (
+	metricSubsystem     = "msgraph"
+	metricsLabelStatus  = "status"
+	metricsLabelsMethod = "method"
+)
+
+func newMetrics() *clientMetrics {
+	return &clientMetrics{
+		requestTotal: prometheus.NewCounterVec(prometheus.CounterOpts{
+			Namespace: teleport.MetricNamespace,
+			Subsystem: metricSubsystem,
+			Name:      "request_total",
+			Help:      "Total number of requests made to MS Graph",
+		}, []string{metricsLabelsMethod, metricsLabelStatus}),
+		requestDuration: prometheus.NewHistogramVec(prometheus.HistogramOpts{
+			Namespace: teleport.MetricNamespace,
+			Subsystem: metricSubsystem,
+			Name:      "request_duration_seconds",
+			Help:      "Request to MS Graph duration in seconds.",
+		}, []string{metricsLabelsMethod}),
+	}
+}
+
+func (metrics *clientMetrics) register(r prometheus.Registerer) error {
+	return trace.NewAggregate(
+		r.Register(metrics.requestTotal),
+		r.Register(metrics.requestDuration),
+	)
+}