WIP: SignalManager API to install signal stacks.

Makefile

@@ -43,7 +43,7 @@ check-jq:
 	jq --version > /dev/null
 
 .PHONY: docs
-docs: $(CRDOC)
+docs: $(CRDOC) generate-crds
 	mkdir -p docs
 	$(CRDOC) --resources deploy/crds/common --output docs/api.md
 

bundle/manifests/observability-operator.clusterserviceversion.yaml

@@ -41,8 +41,8 @@ metadata:
     capabilities: Basic Install
     categories: Monitoring
     certified: "false"
-    containerImage: observability-operator:1.0.0
-    createdAt: "2025-03-03T08:38:22Z"
+    containerImage: grover.home:5000/alanconway/observability-operator:3.3.3
+    createdAt: "2025-04-10T17:19:16Z"
     description: A Go based Kubernetes operator to setup and manage highly available
       Monitoring Stack using Prometheus, Alertmanager and Thanos Querier.
     operatorframework.io/cluster-monitoring: "true"
@@ -58,7 +58,7 @@ metadata:
       ]
     operators.operatorframework.io/project_layout: unknown
     repository: https://github.com/rhobs/observability-operator
-  name: observability-operator.v1.0.0
+  name: observability-operator.v3.3.3
   namespace: placeholder
 spec:
   apiservicedefinitions: {}
@@ -511,23 +511,33 @@ spec:
         - apiGroups:
           - observability.openshift.io
           resources:
-          - uiplugins
+          - signalmanagers
           verbs:
-          - create
-          - delete
           - get
           - list
-          - patch
-          - update
           - watch
         - apiGroups:
           - observability.openshift.io
           resources:
+          - signalmanagers/finalizers
+          - signalmanagers/status
           - uiplugins/finalizers
           - uiplugins/status
           verbs:
           - get
           - update
+        - apiGroups:
+          - observability.openshift.io
+          resources:
+          - uiplugins
+          verbs:
+          - create
+          - delete
+          - get
+          - list
+          - patch
+          - update
+          - watch
         - apiGroups:
           - operator.openshift.io
           resources:
@@ -917,7 +927,7 @@ spec:
                   valueFrom:
                     fieldRef:
                       fieldPath: metadata.namespace
-                image: observability-operator:1.0.0
+                image: grover.home:5000/alanconway/observability-operator:3.3.3
                 imagePullPolicy: Always
                 livenessProbe:
                   httpGet:
@@ -1045,7 +1055,7 @@ spec:
   maturity: alpha
   provider:
     name: Red Hat
-  version: 1.0.0
+  version: 3.3.3
   webhookdefinitions:
   - admissionReviewVersions:
     - v1

cmd/operator/main.go

@@ -40,13 +40,13 @@ var defaultImages = map[string]string{
 	"alertmanager":               "",
 	"thanos":                     obopo.DefaultThanosImage,
 	"ui-dashboards":              "quay.io/openshift-observability-ui/console-dashboards-plugin:v0.4.0",
-	"ui-troubleshooting-panel":   "quay.io/openshift-observability-ui/troubleshooting-panel-console-plugin:v0.4.0",
+	"ui-troubleshooting-panel":   "grover.home:5000/alanconway/troubleshooting-panel-console-plugin:latest",
 	"ui-distributed-tracing-pf4": "quay.io/openshift-observability-ui/distributed-tracing-console-plugin:v0.3.0",
 	"ui-distributed-tracing":     "quay.io/openshift-observability-ui/distributed-tracing-console-plugin:v0.4.0",
 	"ui-logging-pf4":             "quay.io/openshift-observability-ui/logging-view-plugin:v6.0.0",
 	"ui-logging":                 "quay.io/openshift-observability-ui/logging-view-plugin:v6.1.0",
 	"ui-monitoring":              "quay.io/openshift-observability-ui/monitoring-console-plugin:release-coo-1.1",
-	"korrel8r":                   "quay.io/korrel8r/korrel8r:0.7.4",
+	"korrel8r":                   "grover.home:5000/alanconway/korrel8r:latest",
 	"health-analyzer":            "quay.io/openshiftanalytics/cluster-health-analyzer:v0.4.0",
 	"perses":                     "quay.io/persesdev/perses:v0.50.1",
 }
@@ -95,7 +95,7 @@ func main() {
 	flag.StringVar(&metricsAddr, "metrics-bind-address", ":8080", "The address the metric endpoint binds to.")
 	flag.StringVar(&healthProbeAddr, "health-probe-bind-address", ":8081", "The address the health probe endpoint binds to.")
 	flag.Var(images, "images", fmt.Sprintf("Full images refs to use for containers managed by the operator. E.g thanos=quay.io/thanos/thanos:v0.33.0. Images used are %v", imagesUsed()))
-	flag.BoolVar(&openShiftEnabled, "openshift.enabled", false, "Enable OpenShift specific features such as Console Plugins.")
+	flag.BoolVar(&openShiftEnabled, "openshift.enabled", true, "Enable OpenShift specific features such as Console Plugins.")
 
 	opts := zap.Options{
 		Development: true,

deploy/crds/common/observability.openshift.io_configs.yaml

@@ -0,0 +1,94 @@
+---
+apiVersion: apiextensions.k8s.io/v1
+kind: CustomResourceDefinition
+metadata:
+  annotations:
+    controller-gen.kubebuilder.io/version: v0.16.2
+  name: configs.observability.openshift.io
+spec:
+  group: observability.openshift.io
+  names:
+    kind: Config
+    listKind: ConfigList
+    plural: configs
+    singular: config
+  scope: Cluster
+  versions:
+  - name: v1alpha1
+    schema:
+      openAPIV3Schema:
+        description: |-
+          Config defines an observability configuration that can deploy operators and resources for
+          observability signal collectors and stores.
+        properties:
+          apiVersion:
+            description: |-
+              APIVersion defines the versioned schema of this representation of an object.
+              Servers should convert recognized schemas to the latest internal value, and
+              may reject unrecognized values.
+              More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
+            type: string
+          kind:
+            description: |-
+              Kind is a string value representing the REST resource this object represents.
+              Servers may infer this from the endpoint the client submits requests to.
+              Cannot be updated.
+              In CamelCase.
+              More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
+            type: string
+          metadata:
+            type: object
+          spec:
+            description: ConfigSpec specifies what to install.
+            properties:
+              defaultInstall:
+                description: |-
+                  DefaultInstall is the default install type for signals that are not listed or
+                  are listed without an `install` field.
+                  For example, `{ defaultinstall: Default }` with no `signals` field installs all signal
+                  types with default settings.
+                type: string
+              installDefinitions:
+                items:
+                  description: InstallDefinitionSpec defines a new installation type.
+                  properties:
+                    configMap:
+                      description: |-
+                        ConfigMap contains deployment bundles for the install type, with key=signal type.
+                        FIXME: which way around? install < signal or signal < install
+                      type: string
+                    install:
+                      description: Name of the installation type.
+                      type: string
+                  type: object
+                type: array
+              signals:
+                description: Signals specifies what to install for each signal type.
+                items:
+                  properties:
+                    installType:
+                      description: Install type for this signal. Optional, if absent
+                        use ..defaultInstall
+                      type: string
+                    namespace:
+                      description: |-
+                        Namespace to install to. Optional, each signal type has a default namespace.
+                        A signal can be listed multiple times with different `namespace` values,
+                        to install in multiple namespaces.
+                      type: string
+                    signal:
+                      description: Name of this signal type.
+                      type: string
+                  required:
+                  - signal
+                  type: object
+                type: array
+            type: object
+          status:
+            description: FIXME Status
+            type: object
+        type: object
+    served: true
+    storage: true
+    subresources:
+      status: {}

deploy/crds/common/observability.openshift.io_signalmanagers.yaml

@@ -0,0 +1,135 @@
+---
+apiVersion: apiextensions.k8s.io/v1
+kind: CustomResourceDefinition
+metadata:
+  annotations:
+    controller-gen.kubebuilder.io/version: v0.16.2
+  name: signalmanagers.observability.openshift.io
+spec:
+  group: observability.openshift.io
+  names:
+    kind: SignalManager
+    listKind: SignalManagerList
+    plural: signalmanagers
+    singular: signalmanager
+  scope: Cluster
+  versions:
+  - name: v1alpha1
+    schema:
+      openAPIV3Schema:
+        description: "SignalManager is a custom resource to enable observability in
+          the cluster.\n\nEach type of observability signal (logs, metrics, network
+          events, ...) requires operators to be\ninstalled and resources created to
+          configure collection, processing, and storage of signal data.\n\nThe SignalManager
+          automatically installs the operators, custom resource definitions, and\nresources
+          to enable all the desired observability signals in a cluster with default\nconfigurations.\n\nThis
+          means you can get observability up and running quickly and easily,\nbut
+          still customize the details if and when you need to.\n\n## Pattern\n\nA
+          \"Pattern\" is a named set of configurations for each of the observability
+          signals.\nChoosing a pattern automatically installs required operators (if
+          needed) _and_ creates\nworking resources so you have complete, working,
+          observability stacks.\n\nThe following patterns are always available, others
+          may be made available.\n\n  - Default:\n    Installs operators and resources
+          suitable for the most common use cases.\n    The operator owns and manages
+          the resources, and keeps them in the default state.\n  - Custom:\n    Installs
+          operators, but does not create any live resources.\n    The user can create
+          customized resources, they will not be modified by this operator.\n  - Disabled:
+          Do not install any operators, resource definitions, or resources.\n\nCustom
+          patterns can be defined in `spec.patterns`.\n\n## Examples\n\nEnable all
+          observability components with default settings.\n\n\tkind: SignalManager\n\tspec:\n\t
+          \ pattern: Default\n\nDisable all observability components except for logging.\n\n\tkind:
+          SignalManager\n\tspec:\n\t  pattern: Disabled\n\t  signals:\n\t    name:
+          Log\n\t    pattern: Default\n\nEnable most components with defaults, install
+          the logging operators,\nbut use custom logging resources (created separately)\n\n\tkind:
+          SignalManager\n\tspec:\n\t  pattern: Default\n\t  signals:\n\t    name:
+          log\n\t    pattern: Custom\n\n## Lifecycle and ownership\n\nOwnership of
+          resources depends on the pattern:\n\n  - None: No operators installed, no
+          resources created or reconciled.\n  - Custom: Operators installed but no
+          resources created. User is free to create resources\n    they are not owned
+          or reconciled by this operator.\n  - Default, or any other defined configuration:\n
+          \   This operator creates, owns, and reconciles resources to keep them consistent
+          with the chosen\n    pattern.\n\nFIXME: Operator may reconcile only part
+          of the resource and allow user to tweak other parts.\nNeeds consideration.
+          COO already uses server-side-apply to do this in some cases.\n\nFIXME: Patterns
+          may need to be \"parameterized\" e.g. with sizing data.\nHow to include
+          such parameters without duplicating existing CRs?\n\nFIXME: Define behavior
+          on spec changes: deleting, re-creating, updating resources.\nChange to Custom
+          should leave resources in place so user can eddit.\nWhat to do on change
+          _from_ Custom?"
+        properties:
+          apiVersion:
+            description: |-
+              APIVersion defines the versioned schema of this representation of an object.
+              Servers should convert recognized schemas to the latest internal value, and
+              may reject unrecognized values.
+              More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
+            type: string
+          kind:
+            description: |-
+              Kind is a string value representing the REST resource this object represents.
+              Servers may infer this from the endpoint the client submits requests to.
+              Cannot be updated.
+              In CamelCase.
+              More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
+            type: string
+          metadata:
+            type: object
+          spec:
+            description: Lists signals and the pattern to deploy them.
+            properties:
+              pattern:
+                description: The default pattern for signals that are not listed or
+                  have no `pattern` field.
+                type: string
+              patterns:
+                description: Patterns is a list of custom pattern definitions.
+                items:
+                  description: |-
+                    PatternSpec defines a custom pattern.
+
+                    on the cluster. Simplest format is a flat YAML file, but we may need more structure
+                    to store kustomize scripts, multi-stage deployments, health checks, metadata etc....
+                    Possible storage formats: ConfigMap, PersistentVolume, container image...
+
+                    Patterns should also be usable directly, without depending on this API.
+                    Preferably using only kubectl and kustomize.
+                  properties:
+                    pattern:
+                      description: Name of the pattern.
+                      type: string
+                  required:
+                  - pattern
+                  type: object
+                type: array
+              signals:
+                description: Signals is a list of signal types with the desired pattern.
+                items:
+                  properties:
+                    name:
+                      description: Signal name
+                      type: string
+                    namespace:
+                      description: |-
+                        Namespace to install to.
+
+                        Optional, each signal type has a default namespace.
+                      type: string
+                    pattern:
+                      description: Pattern for this signal. Optional, if absent use
+                        the 'Default' pattern.
+                      type: string
+                  required:
+                  - name
+                  type: object
+                type: array
+            required:
+            - pattern
+            type: object
+          status:
+            description: Status of the signal manager.
+            type: object
+        type: object
+    served: true
+    storage: true
+    subresources:
+      status: {}

deploy/olm/kustomization.yaml

@@ -12,8 +12,8 @@ resources:
 
 images:
 - name: observability-operator
-  newName: observability-operator
-  newTag: 1.0.0
+  newName: grover.home:5000/alanconway/observability-operator
+  newTag: 3.3.3
 
 patches:
 - patch: |-

deploy/operator/observability-operator-cluster-role.yaml

@@ -226,23 +226,33 @@ rules:
 - apiGroups:
   - observability.openshift.io
   resources:
-  - uiplugins
+  - signalmanagers
   verbs:
-  - create
-  - delete
   - get
   - list
-  - patch
-  - update
   - watch
 - apiGroups:
   - observability.openshift.io
   resources:
+  - signalmanagers/finalizers
+  - signalmanagers/status
   - uiplugins/finalizers
   - uiplugins/status
   verbs:
   - get
   - update
+- apiGroups:
+  - observability.openshift.io
+  resources:
+  - uiplugins
+  verbs:
+  - create
+  - delete
+  - get
+  - list
+  - patch
+  - update
+  - watch
 - apiGroups:
   - operator.openshift.io
   resources: