Merge pull request #211 from tv2-oss/feature/add-metrics

Add metrics

Author: michaelvl

controllers/common.go

@@ -317,6 +317,7 @@ func patchUnstructured(ctx context.Context, r ControllerDynClient, us *unstructu
 
 	force := true
 
+	metricPatchApply.Inc()
 	if namespace != nil {
 		dynamicClient := r.DynamicClient().Resource(*gvr).Namespace(*namespace)
 		_, err = dynamicClient.Patch(ctx, us.GetName(), types.ApplyPatchType, jsonData, metav1.PatchOptions{
@@ -331,5 +332,8 @@ func patchUnstructured(ctx context.Context, r ControllerDynClient, us *unstructu
 		})
 	}
 
+	if err != nil {
+		metricPatchApplyErrs.Inc()
+	}
 	return err
 }

controllers/metrics.go

@@ -0,0 +1,74 @@
+/*
+Copyright 2023 TV 2 DANMARK A/S
+
+Licensed under the Apache License, Version 2.0 (the "License") with the
+following modification to section 6. Trademarks:
+
+Section 6. Trademarks is deleted and replaced by the following wording:
+
+6. Trademarks. This License does not grant permission to use the trademarks and
+trade names of TV 2 DANMARK A/S, including but not limited to the TV 2® logo and
+word mark, except (a) as required for reasonable and customary use in describing
+the origin of the Work, e.g. as described in section 4(c) of the License, and
+(b) to reproduce the content of the NOTICE file. Any reference to the Licensor
+must be made by making a reference to "TV 2 DANMARK A/S", written in capitalized
+letters as in this example, unless the format in which the reference is made,
+requires lower case letters.
+
+You may not use this software except in compliance with the License and the
+modifications set out above.
+
+You may obtain a copy of the license at:
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+*/
+
+package controllers
+
+import (
+	"github.com/prometheus/client_golang/prometheus"
+	"sigs.k8s.io/controller-runtime/pkg/metrics"
+)
+
+var (
+	metricPatchApply = prometheus.NewCounter(
+		prometheus.CounterOpts{
+			Name: "bifrost_patchapply_total",
+			Help: "Number of server-side patch operations",
+		},
+	)
+	metricPatchApplyErrs = prometheus.NewCounter(
+		prometheus.CounterOpts{
+			Name: "bifrost_patchapply_errors_total",
+			Help: "Number of server-side patch errors",
+		},
+	)
+	metricTemplateErrs = prometheus.NewCounter(
+		prometheus.CounterOpts{
+			Name: "bifrost_template_errors_total",
+			Help: "Number of template render errors",
+		},
+	)
+	metricTemplateParseErrs = prometheus.NewCounter(
+		prometheus.CounterOpts{
+			Name: "bifrost_template_parse_errors_total",
+			Help: "Number of template parse errors",
+		},
+	)
+	metricResourceGet = prometheus.NewCounter(
+		prometheus.CounterOpts{
+			Name: "bifrost_resource_get_total",
+			Help: "Number of resources fetched to use as dependency in templates",
+		},
+	)
+)
+
+func init() {
+	metrics.Registry.MustRegister(metricPatchApply, metricPatchApplyErrs, metricTemplateErrs, metricResourceGet)
+}

controllers/templating.go

@@ -131,6 +131,7 @@ func parseTemplates(resourceTemplates map[string]string) ([]*ResourceTemplateSta
 		r.StringTemplate = tmpl
 		r.Template, err = parseSingleTemplate(tmplKey, tmpl)
 		if err != nil {
+			metricTemplateParseErrs.Inc()
 			return nil, fmt.Errorf("cannot parse template %q: %w", tmplKey, err)
 		}
 		r.NewResources = make([]ResourceComposite, 0)
@@ -165,6 +166,7 @@ func renderTemplates(ctx context.Context, r ControllerDynClient, parent metav1.O
 					// FIXME: These are convenient, but we should have a better logging design, i.e. it should be possible to enable rendering errors only
 					fmt.Printf("Template:\n%s\n", tmpl.StringTemplate)
 					fmt.Printf("Template values:\n%+v\n", values)
+					metricTemplateErrs.Inc()
 				}
 				continue
 			}
@@ -179,6 +181,7 @@ func renderTemplates(ctx context.Context, r ControllerDynClient, parent metav1.O
 				} else {
 					dynamicClient = r.DynamicClient().Resource(*res.GVR)
 				}
+				metricResourceGet.Inc()
 				res.Current, err = dynamicClient.Get(ctx, res.Rendered.GetName(), metav1.GetOptions{})
 				if err != nil {
 					logger.Error(err, "cannot get current resource", "templateName", tmpl.TemplateName, "resIdx", resIdx)

doc/installing.md

@@ -43,11 +43,21 @@ GatewayClassBlueprints](../blueprints/README.md).
 
 ## Metrics and Observability
 
-The controller provides [standard controller
+The controller provides the following Prometheus/OpenMetrics metrics:
+
+| Metric | Type | Description |
+| ------------- | ------------- |
+| `bifrost_patchapply_total` | Counter | Number of server-side patch operations |
+| `bifrost_patchapply_errors_total` | Counter | Number of server-side patch errors |
+| `bifrost_template_errors_total` | Counter | Number of template render errors |
+| `bifrost_template_parse_errors_total` | Counter | Number of template parse errors |
+| `bifrost_resource_get_total` | Counter | Number of resources fetched to use as dependency in templates |
+
+Additionally the controller provides [standard controller
 metrics](https://book.kubebuilder.io/reference/metrics-reference.html)
 and the Helm chart provides a
 [ServiceMonitor](https://github.com/prometheus-operator/prometheus-operator/blob/main/Documentation/api.md#monitoring.coreos.com/v1.ServiceMonitor)
-for integration with systems that understand these.
+for integration with systems that understand this CRD.
 
 Observability of Gateway-API resources is possible through [Custom
 Resource State