Feature/observability: scripts and doc updates

README.md

@@ -30,7 +30,7 @@ Apart from the above arrangement, the following system modules/pods are part of
 
 This section describes the nested topology design implemented by this solution.
 
-![alt text](architecture/nested-topology-hld-envoy.png "Nested Toplogy")
+![alt text](architecture/nested-topology-hld.png "Nested Toplogy")
 
 At the core of the nested topology design, we have reverse proxies which broker the connections between each hypothetical ISA-95 level (Level 2,3,4 in this instance). These proxies prevent workloads and Arc agents running at lower levels from connecting to the outside world directly, allowing the traffic to be managed or controlled via proxy configuration at each level. Currently, data plane is traversing layers directly between brokers, and we are evaluating an improvement to force this communication to pass through the proxy transparently.
 Proxying of allowed URI calls from the lower L2 and L3 levels for the AKS host nodes (kubelet, containerd) is implemented using a DNS Server override in each lower Virtual Network.
@@ -68,6 +68,10 @@ Workloads exchange messages locally on the same network layer and Kubernetes clu
 
 For more information about MQTT broker choice and comparison, please see [MQTT Broker for Data Communication Between Workloads and Between Network Layers](/docs/mqttbroker.md).
 
+### Monitoring and Observability
+
+Gathering diverse signals from sources such as operating systems, data components, custom workloads, and the Kubernetes platform itself, as well as analyzing these is discussed in a separate document [Observability for Distributed Edge](./docs/observability.md).
+
 ## Solution Deployment
 
 ![alt text](architecture/deployment-hld.png "Deployment Strategy")

deployment/bicep/core-infra-vnet.bicep

@@ -81,6 +81,9 @@ param aksObjectId string
 @description('Wether to close down outbound internet access')
 param closeOutboundInternetAccess bool = false
 
+@description('Provision monitoring')
+param provisionMonitoring bool = false
+
 var applicationNameWithoutDashes = replace(applicationName, '-', '')
 var aksName = take('aks-${applicationNameWithoutDashes}', 20)
 var resourceGroupName = applicationName
@@ -131,6 +134,16 @@ module downstreamvnetpeering 'modules/vnetpeering.bicep' = if (!empty(remoteVnet
   ]
 }
 
+module monitoring 'modules/loganalytics.bicep' = if (provisionMonitoring) {
+  scope: resourceGroup(rg.name)
+  name: 'monitoringDeployment'
+  params: {
+    workspaceAccountName: applicationName
+    monitorAccountLocation: location
+  }
+}
+
 output aksName string = aksName
 output aksResourceGroup string = resourceGroupName
 output subnetId string = vnet.outputs.subnetId
+output appInsightsInstrumentationKey string = provisionMonitoring ? monitoring.outputs.instrumentationKey : ''

deployment/bicep/modules/azurestorage.bicep

@@ -4,7 +4,7 @@
 // ------------------------------------------------------------
 @minLength(3)
 @maxLength(24)
-@description('Azure Stroage Account name which is not already in use.')
+@description('Azure Storage Account name which is not already in use.')
 param storageAccountName string
 
 @description('Storage account location')

deployment/bicep/modules/loganalytics.bicep

@@ -0,0 +1,50 @@
+// ------------------------------------------------------------
+//  Copyright (c) Microsoft Corporation.  All rights reserved.
+//  Licensed under the MIT License (MIT). See License.txt in the repo root for license information.
+// ------------------------------------------------------------
+@minLength(3)
+@maxLength(24)
+@description('Azure Log Analytics and App Insights Account name which is not already in use.')
+param workspaceAccountName string
+
+@description('Azure Log Analytics account location')
+@maxLength(20)
+param monitorAccountLocation string = resourceGroup().location
+
+resource logAnalytics 'Microsoft.OperationalInsights/workspaces@2022-10-01' = {
+  name: workspaceAccountName
+  location: monitorAccountLocation
+  properties: {
+    sku: {
+      name: 'PerGB2018'
+    }
+    retentionInDays: 30
+    features: {
+      enableLogAccessUsingOnlyResourcePermissions: true
+    }
+    workspaceCapping: {
+      dailyQuotaGb: -1
+    }
+    publicNetworkAccessForIngestion: 'Enabled'
+    publicNetworkAccessForQuery: 'Enabled'
+  }
+}
+
+resource appInsightsComponent 'microsoft.insights/components@2020-02-02' = {
+  name: workspaceAccountName
+  location: monitorAccountLocation
+  kind: 'web'
+  properties: {
+    Application_Type: 'web'
+    Flow_Type: 'Bluefield'
+    Request_Source: 'rest'
+    RetentionInDays: 90
+    WorkspaceResourceId: logAnalytics.id
+    IngestionMode: 'LogAnalytics'
+    publicNetworkAccessForIngestion: 'Enabled'
+    publicNetworkAccessForQuery: 'Enabled'
+  }
+}
+
+output instrumentationKey string = appInsightsComponent.properties.InstrumentationKey
+

deployment/bicep/modules/vnet.bicep

@@ -31,8 +31,8 @@ param closeOutboundInternetAccess bool = false
 var subnetName = aksName
 var subnetNsgName = aksName
 
-var arrayBasicRules = [ allowProxyInboundSecurityRule, allowMqttSslInboundSecurityRule ]
-var arrayBaseAndLockRules = [ allowProxyInboundSecurityRule, allowMqttSslInboundSecurityRule, allowK8ApiHTTPSOutbound, allowK8ApiUdpOutbound, allowTagAks9000Outbound, allowTagFrontDoorFirstParty, allowTagMcr, denyOutboundInternetAccessSecurityRule ]
+var arrayBasicRules = [ allowProxyInboundSecurityRule, allowMqttSslInboundSecurityRule, allowOtelGrpcInboundSecurityRule ]
+var arrayBaseAndLockRules = [ allowProxyInboundSecurityRule, allowMqttSslInboundSecurityRule, allowOtelGrpcInboundSecurityRule, allowK8ApiHTTPSOutbound, allowK8ApiUdpOutbound, allowTagAks9000Outbound, allowTagFrontDoorFirstParty, allowTagMcr, denyOutboundInternetAccessSecurityRule ]
 
 // TODO: We need to do this is nested manner e.g. use parent vnet/subnet if this is nested vnet/subnet creation.
 var allowProxyInboundSecurityRule = {
@@ -41,15 +41,14 @@ var allowProxyInboundSecurityRule = {
     priority: 1010
     access: 'Allow'
     direction: 'Inbound'
-    destinationPortRange: '443'
+    destinationPortRanges: ['443', '8084']
     protocol: 'Tcp'
     sourcePortRange: '*'
     sourceAddressPrefix: 'VirtualNetwork'
     destinationAddressPrefix: 'VirtualNetwork'
   }
 }
 
-// TODO: potentially remove this if going through proxy, for now setup for testing MQTT bridging
 var allowMqttSslInboundSecurityRule = {
   name: 'AllowMqttSsl'
   properties: {
@@ -64,6 +63,20 @@ var allowMqttSslInboundSecurityRule = {
   }
 }
 
+var allowOtelGrpcInboundSecurityRule = {
+  name: 'AllowOtelGrpc'
+  properties: {
+    priority: 1030
+    access: 'Allow'
+    direction: 'Inbound'
+    destinationPortRange: '4318'
+    protocol: 'Tcp'
+    sourcePortRange: '*'
+    sourceAddressPrefix: 'VirtualNetwork'
+    destinationAddressPrefix: 'VirtualNetwork'
+  }
+}
+
 var allowK8ApiHTTPSOutbound = {
   name: 'AllowK8ApiHTTPSOutbound'
   properties: {

deployment/build-and-deploy-images.ps1

@@ -14,7 +14,11 @@ Param(
     # leave empty if both workloads are deployed on single cluster L4
     [string]
     [Parameter(mandatory=$false)]
-    $L2ResourceGroupName
+    $L2ResourceGroupName,
+
+    [Parameter(Mandatory = $false)]
+    [bool]
+    $SetupObservability = $true
 )
 
 if(!$env:RESOURCEGROUPNAME -and !$AppResourceGroupName)
@@ -56,6 +60,8 @@ Set-Location -Path $deploymentDir
 
 Write-Title("Upgrade/Install Pod/Containers with Helm charts in Cluster L4")
 $datagatewaymoduleimage = $acrName + ".azurecr.io/datagatewaymodule:" + $deploymentId
+$observabilityString = ($SetupObservability -eq $true) ? "true" : "false"
+$samplingRate = ($SetupObservability -eq $true) ? "1" : "0" # in development we set to 1, in prod should be 0.0001 or similar, 0 turns off observability
 
 # ----- Get Cluster Credentials for L4 layer
 Write-Title("Get AKS Credentials L4 Layer")
@@ -68,6 +74,8 @@ az aks get-credentials `
 
 helm upgrade iot-edge-l4 ./helm/iot-edge-l4 `
     --set-string images.datagatewaymodule="$datagatewaymoduleimage" `
+    --set-string observability.samplingRate="$samplingRate" `
+    --set observability.enabled=$observabilityString `
     --namespace $appKubernetesNamespace `
     --reuse-values `
     --install
@@ -96,6 +104,8 @@ helm upgrade iot-edge-l2 ./helm/iot-edge-l2 `
     --set-string images.simulatedtemperaturesensormodule="$simtempimage" `
     --set-string images.opcplcmodule="$opcplcimage" `
     --set-string images.opcpublishermodule="$opcpublisherimage" `
+    --set observability.enabled=$observabilityString `
+    --set-string observability.samplingRate="$samplingRate" `
     --reuse-values `
     --namespace $appKubernetesNamespace `
     --install

deployment/deploy-az-demo-bootstrapper.ps1

@@ -13,7 +13,11 @@ Param(
 
     [string]
     [Parameter(mandatory=$false)]
-    $Location = 'westeurope'    
+    $Location = 'westeurope',
+
+    [Parameter(Mandatory = $false)]
+    [bool]
+    $SetupObservability = $true
 )
 
 mkdir -p modules
@@ -45,11 +49,12 @@ Invoke-WebRequest -Uri "$baseLocation/deployment/bicep/modules/azurestorage.bice
 Invoke-WebRequest -Uri "$baseLocation/deployment/bicep/modules/eventhub.bicep" -OutFile "./bicep/modules/eventhub.bicep"
 Invoke-WebRequest -Uri "$baseLocation/deployment/bicep/modules/vnet.bicep" -OutFile "./bicep/modules/vnet.bicep"
 Invoke-WebRequest -Uri "$baseLocation/deployment/bicep/modules/vnetpeering.bicep" -OutFile "./bicep/modules/vnetpeering.bicep"
+Invoke-WebRequest -Uri "$baseLocation/deployment/bicep/modules/loganalytics.bicep" -OutFile "./bicep/modules/loganalytics.bicep"
 
 # Deploy 3 core infrastructure layers i.e. L4, L3, L2, replicating 3 levels of Purdue network topology.
-$l4LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ApplicationName ($ApplicationName + "L4") -VnetAddressPrefix "172.16.0.0/16" -SubnetAddressPrefix "172.16.0.0/18" -SetupArc $true -Location $Location
-$l3LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ParentConfig $l4LevelCoreInfra -ApplicationName ($ApplicationName + "L3") -VnetAddressPrefix "172.18.0.0/16" -SubnetAddressPrefix "172.18.0.0/18" -SetupArc $true -Location $Location
-$l2LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ParentConfig $l3LevelCoreInfra -ApplicationName ($ApplicationName + "L2") -VnetAddressPrefix "172.20.0.0/16" -SubnetAddressPrefix "172.20.0.0/18" -SetupArc $true -Location $Location
+$l4LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ApplicationName ($ApplicationName + "L4") -VnetAddressPrefix "172.16.0.0/16" -SubnetAddressPrefix "172.16.0.0/18" -SetupArc $true -Location $Location -SetupObservability $SetupObservability
+$l3LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ParentConfig $l4LevelCoreInfra -ApplicationName ($ApplicationName + "L3") -VnetAddressPrefix "172.18.0.0/16" -SubnetAddressPrefix "172.18.0.0/18" -SetupArc $true -Location $Location -SetupObservability $SetupObservability
+$l2LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ParentConfig $l3LevelCoreInfra -ApplicationName ($ApplicationName + "L2") -VnetAddressPrefix "172.20.0.0/16" -SubnetAddressPrefix "172.20.0.0/18" -SetupArc $true -Location $Location -SetupObservability $SetupObservability
 
 # Deploy core platform layer (Dapr on L4 and L2, Mosquitto broker bridging on L2, L3 and L4).
 $l4CorePlatform = ./deploy-core-platform.ps1 -AksClusterName $l4LevelCoreInfra.AksClusterName -AksClusterResourceGroupName $l4LevelCoreInfra.AksClusterResourceGroupName -DeployDapr $true -MosquittoParentConfig $null

deployment/deploy-az-dev-bootstrapper.ps1

@@ -9,7 +9,15 @@ Param(
 
     [string]
     [Parameter(mandatory=$false)]
-    $Location = 'westeurope'
+    $Location = 'westeurope',
+
+    [Parameter(Mandatory = $false)]
+    [bool]
+    $SetupObservability = $true,
+
+    [Parameter(Mandatory = $false)]
+    [bool]
+    $SetupArc = $false
 )
 
 # Import text utilities module.
@@ -28,48 +36,50 @@ $Location = Get-AzShortRegion($Location)
 
 # 1. Deploy core infrastructure (AKS clusters, VNET)
 
-$l4LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ApplicationName ($ApplicationName + "L4") -VnetAddressPrefix "172.16.0.0/16" -SubnetAddressPrefix "172.16.0.0/18" -SetupArc $false -Location $Location
-$l3LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ParentConfig $l4LevelCoreInfra -ApplicationName ($ApplicationName + "L3") -VnetAddressPrefix "172.18.0.0/16" -SubnetAddressPrefix "172.18.0.0/18" -SetupArc $false -Location $Location
-$l2LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ParentConfig $l3LevelCoreInfra -ApplicationName ($ApplicationName + "L2") -VnetAddressPrefix "172.20.0.0/16" -SubnetAddressPrefix "172.20.0.0/18" -SetupArc $false -Location $Location
+$l4LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ApplicationName ($ApplicationName + "L4") -VnetAddressPrefix "172.16.0.0/16" -SubnetAddressPrefix "172.16.0.0/18" -SetupArc $SetupArc -Location $Location -SetupObservability $SetupObservability
+$l3LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ParentConfig $l4LevelCoreInfra -ApplicationName ($ApplicationName + "L3") -VnetAddressPrefix "172.18.0.0/16" -SubnetAddressPrefix "172.18.0.0/18" -SetupArc $SetupArc -Location $Location -SetupObservability $SetupObservability
+$l2LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ParentConfig $l3LevelCoreInfra -ApplicationName ($ApplicationName + "L2") -VnetAddressPrefix "172.20.0.0/16" -SubnetAddressPrefix "172.20.0.0/18" -SetupArc $SetupArc -Location $Location -SetupObservability $SetupObservability
 
-# # 2. Deploy core platform in each layer (Dapr, Mosquitto and bridging).
-$l4CorePlatform = ./deploy-core-platform.ps1 -AksClusterName $l4LevelCoreInfra.AksClusterName -AksClusterResourceGroupName $l4LevelCoreInfra.AksClusterResourceGroupName -DeployDapr $true -MosquittoParentConfig $null -ArcEnabled $false
-$l3CorePlatform = ./deploy-core-platform.ps1 -AksClusterName $l3LevelCoreInfra.AksClusterName -AksClusterResourceGroupName $l3LevelCoreInfra.AksClusterResourceGroupName -MosquittoParentConfig $l4CorePlatform -ArcEnabled $false
-./deploy-core-platform.ps1 -AksClusterName $l2LevelCoreInfra.AksClusterName -AksClusterResourceGroupName $l2LevelCoreInfra.AksClusterResourceGroupName -DeployDapr $true -MosquittoParentConfig $l3CorePlatform -ArcEnabled $false
+# 2. Deploy core platform in each layer (Dapr, Mosquitto and bridging).
+$l4CorePlatform = ./deploy-core-platform.ps1 -AksClusterName $l4LevelCoreInfra.AksClusterName -AksClusterResourceGroupName $l4LevelCoreInfra.AksClusterResourceGroupName -DeployDapr $true -MosquittoParentConfig $null -ArcEnabled $SetupArc
+$l3CorePlatform = ./deploy-core-platform.ps1 -AksClusterName $l3LevelCoreInfra.AksClusterName -AksClusterResourceGroupName $l3LevelCoreInfra.AksClusterResourceGroupName -MosquittoParentConfig $l4CorePlatform -ArcEnabled $SetupArc
+./deploy-core-platform.ps1 -AksClusterName $l2LevelCoreInfra.AksClusterName -AksClusterResourceGroupName $l2LevelCoreInfra.AksClusterResourceGroupName -DeployDapr $true -MosquittoParentConfig $l3CorePlatform -ArcEnabled $SetupArc
 
 # 3. Deploy app resources in Azure, build images and deploy helm on level L4 and L2.
 $l4AppConfig = ./deploy-dev-app-l4.ps1 -ApplicationName $ApplicationName `
     -AksClusterResourceGroupName $l4LevelCoreInfra.AksClusterResourceGroupName `
     -AksClusterName $l4LevelCoreInfra.AksClusterName -AksServicePrincipalName ($ApplicationName + "L4") `
-    -Location $Location
+    -Location $Location `
+    -SetupObservability $SetupObservability
 
 # Note currently for developer flow we need Azure Container Registry deployed by L4 (via L4AppConfig). 
 ./deploy-dev-app-l2.ps1  -ApplicationName $ApplicationName `
     -AksClusterName $l2LevelCoreInfra.AksClusterName `
     -AksClusterResourceGroupName $l2LevelCoreInfra.AksClusterResourceGroupName `
     -AksServicePrincipalName ($ApplicationName + "L2") `
-    -L4AppConfig $l4AppConfig
+    -L4AppConfig $l4AppConfig `
+    -SetupObservability $SetupObservability
 
 # # --- Deploying just a single layer: comment above block and uncomment below:
 
-# $l4LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ApplicationName ($ApplicationName + "L4") -VnetAddressPrefix "172.16.0.0/16" -SubnetAddressPrefix "172.16.0.0/18" -SetupArc $false -Location $Location
+# $l4LevelCoreInfra = ./deploy-core-infrastructure.ps1 -ApplicationName ($ApplicationName + "L4") -VnetAddressPrefix "172.16.0.0/16" -SubnetAddressPrefix "172.16.0.0/18" -SetupArc $false -Location $Location $SetupObservability
 
 # ./deploy-core-platform.ps1 -AksClusterName $l4LevelCoreInfra.AksClusterName `
 #     -AksClusterResourceGroupName $l4LevelCoreInfra.AksClusterResourceGroupName `
-#     -DeployDapr $true -MosquittoParentConfig $null -ArcEnabled $false
+#     -DeployDapr $true -MosquittoParentConfig $null -ArcEnabled $SetupArc
 
 # $l4AppConfig = ./deploy-dev-app-l4.ps1 -ApplicationName $ApplicationName `
 #     -AksClusterResourceGroupName $l4LevelCoreInfra.AksClusterResourceGroupName `
 #     -AksClusterName $l4LevelCoreInfra.AksClusterName `
 #     -AksServicePrincipalName ($ApplicationName + "L4") `
-#     -Location $Location
+#     -Location $Location -SetupObservability $SetupObservability
 
 # # when deploying L2 workload on single cluster in L4, passing in parameters pointing to L4 is intentional
 # ./deploy-dev-app-l2.ps1  -ApplicationName $ApplicationName `
 #     -AksClusterName $l4LevelCoreInfra.AksClusterName `
 #     -AksClusterResourceGroupName $l4LevelCoreInfra.AksClusterResourceGroupName `
 #     -AksServicePrincipalName ($ApplicationName + "L4") `
-#     -L4AppConfig $l4AppConfig
+#     -L4AppConfig $l4AppConfig -SetupObservability $SetupObservability
 # #----------------
 
 $runningTime = New-TimeSpan -Start $startTime