I have prepared an ARM template for deploying an Azure Eventhub instance and wonder how to access the both connection keys for returning them as output?
I would like to return a string in the form:
Endpoint=sb://my-eventhub.servicebus.windows.net/;SharedAccessKeyName=RootManageSharedAccessKey;SharedAccessKey=ojZMQcJD7uYifxJyGeXG6tNDdZyaC1/h5tmX6ODVfmY=
Here is my current template:
{
"$schema": "https://schema.management.azure.com/schemas/2019-04-01/deploymentTemplate.json#",
"contentVersion": "1.0.0.0",
"parameters": {
"clusterName": {
"type": "string",
"defaultValue": "eventhub",
"metadata": {
"description": "Name for the Event Hub cluster."
}
},
"namespaceName": {
"type": "string",
"defaultValue": "namespace",
"metadata": {
"description": "Name for the Namespace to be created in cluster."
}
}
},
"variables": {
"clusterName": "[concat(resourceGroup().name, '-', parameters('clusterName'))]",
"namespaceName": "[concat(resourceGroup().name, '-', parameters('namespaceName'))]"
},
"outputs": {
"MyClusterName": {
"type": "string",
"value": "[variables('clusterName')]"
},
"PrimaryConnectionString": {
"type": "string",
"value": "WHAT TO USE HERE PLEASE?"
},
"SecondaryConnectionString": {
"type": "string",
"value": "WHAT TO USE HERE PLEASE?"
}
},
"resources": [
{
"type": "Microsoft.EventHub/clusters",
"apiVersion": "2018-01-01-preview",
"name": "[variables('clusterName')]",
"location": "[resourceGroup().location]",
"sku": {
"name": "Dedicated",
"capacity": 1
}
},
{
"type": "Microsoft.EventHub/namespaces",
"apiVersion": "2018-01-01-preview",
"name": "[variables('namespaceName')]",
"location": "[resourceGroup().location]",
"dependsOn": [
"[resourceId('Microsoft.EventHub/clusters', variables('clusterName'))]"
],
"sku": {
"name": "Standard",
"tier": "Standard",
"capacity": 1
},
"properties": {
"isAutoInflateEnabled": false,
"maximumThroughputUnits": 0,
"clusterArmId": "[resourceId('Microsoft.EventHub/clusters', variables('clusterName'))]"
}
}
]
}
I have tried the following:
"value": "[listKeys(resourceId(concat('Microsoft.ServiceBus/namespaces/AuthorizationRules'), variables('namespaceName'), 'RootManageSharedAccessKey'),'2018-01-01-preview').primaryConnectionString]"
but get deployment error:
[error]ParentResourceNotFound: Can not perform requested operation on nested resource. Parent resource 'my-rg-namespace' not found.
UPDATE:
The following has worked for me as suggested by Jesse (thank you!):
"variables": {
"clusterName": "[concat(resourceGroup().name, '-', parameters('clusterName'))]",
"namespaceName": "[concat(resourceGroup().name, '-', parameters('namespaceName'))]",
"defaultSASKeyName": "RootManageSharedAccessKey",
"authRuleResourceId": "[resourceId('Microsoft.EventHub/namespaces/authorizationRules', variables('namespaceName'), variables('defaultSASKeyName'))]"
},
"outputs": {
"MyClusterName": {
"type": "string",
"value": "[variables('clusterName')]"
},
"PrimaryConnectionString": {
"type": "string",
"value": "[listkeys(variables('authRuleResourceId'), '2015-08-01').primaryConnectionString]"
},
"SecondaryConnectionString": {
"type": "string",
"value": "[listkeys(variables('authRuleResourceId'), '2015-08-01').secondaryConnectionString]"
}
},
UPDATE 2:
Also, Jesse has noticed that my ARM template is wrong in 2 ways, because it does not create an Event Hub, but a cluster and it is outside my namespace and provided this valuable comment:
The Event Hubs cluster is basically a way of reserving dedicated compute. It's not something that most scenarios need and it is... not cheap. Think of something on the scale of Xbox Live where you're seeing nearly 5 millions of events per second and which have higher performance needs. If you're not looking at that kind of scale or that sensitivity around timing, you probably want to rethink the need for a dedicated cluster.
Normally, you'd just provision an Event Hubs namespace which will use shared infrastructure with certain guarantees to minimize noisy neighbors and similar. This is adequate for the majority of scenarios, even those with high throughput needs. If you're not sure, this is probably the place that you want to start and then upgrade to a dedicated cluster if your needs justify the cost.
An Event Hubs namespace is the container for a set of Event Hub instances grouped together by a unique endpoint. Each Event Hub is made of a set of partitions. When you're publishing or consuming events, the partitions of an Event Hub are where the actual data is. When you're working with one of the SDKs, you'll start by telling it about the endpoint of your namespace and the Event Hub that you're interested in. You'll need a general awareness of partitions, but most of the "Getting Started" scenarios handle that detail for you, as do a fair portion of the real-world ones.... but, the concept is an important one.