Edit

Monitoring Azure Service Bus data reference

  • Article

See Monitoring Azure Service Bus for details on collecting and analyzing monitoring data for Azure Service Bus.

Note

Azure Monitor doesn't include dimensions in the exported metrics data sent to a destination like Azure Storage, Azure Event Hubs, Log Analytics, etc.

Metrics

This section lists all the automatically collected platform metrics collected for Azure Service Bus. The resource provider for these metrics is Microsoft.ServiceBus/namespaces.

Request metrics

Counts the number of data and management operations requests.

Metric Name Exportable via diagnostic settings Unit Aggregation type Description Dimensions
Incoming Requests Yes Count Total The number of requests made to the Service Bus service over a specified period. EntityName
Successful Requests No Count Total The number of successful requests made to the Service Bus service over a specified period. Entity name
OperationResult
Server Errors No Count Total The number of requests not processed because of an error in the Service Bus service over a specified period. Entity name
OperationResult
User Errors No Count Total The number of requests not processed because of user errors over a specified period. Entity name
Throttled Requests No Count Total

The number of requests that were throttled because the usage was exceeded.

MessagingErrorSubCode dimension has the following possible values:

  • CPU: CPU throttling
  • Storage:It indicates throttle because of pending checkpoint operations
  • Namespace:Namespace operations throttling.
  • Unknown: Other resource throttling.

 Entity name
MessagingErrorSubCode
Pending Checkpoint Operations Count No count Average The number of pending checkpoint operations on the namespace. Service starts to throttle when the pending checkpoint count exceeds limit of (500,000 + (500,000 * messaging units)) operations. This metric applies only to namespaces using the premium tier. MessagingErrorSubCode
Server Send Latency No milliseconds Average The time taken by the Service Bus service to complete the request. Entity name

The following two types of errors are classified as user errors:

  1. Client-side errors (In HTTP that would be 400 errors).
  2. Errors that occur while processing messages, such as MessageLockLostException.

Message metrics

Metric Name Exportable via diagnostic settings Unit Aggregation type Description Dimensions
Incoming Messages Yes Count Total The number of events or messages sent to Service Bus over a specified period. For basic and standard tiers, incoming auto-forwarded messages are included in this metric. And, for the premium tier, they aren't included. Entity name
Outgoing Messages Yes Count Total The number of events or messages received from Service Bus over a specified period. The outgoing auto-forwarded messages aren't included in this metric. Entity name
Messages No Count Average Count of messages in a queue/topic. This metric includes messages in all the different states like active, dead-lettered, scheduled, etc. Entity name
Active Messages No Count Average Count of active messages in a queue/topic. Active messages are the messages in the queue or subscription that are in the active state and ready for delivery. The messages are available to be received. Entity name
Dead-lettered messages No Count Average Count of dead-lettered messages in a queue/topic. Entity name
Scheduled messages No Count Average Count of scheduled messages in a queue/topic. Entity name
Completed Messages Yes Count Total The number of messages completed over a specified period. Entity name
Abandoned Messages Yes Count Total The number of messages abandoned over a specified period. Entity name
Size No Bytes Average Size of an entity (queue or topic) in bytes. Entity name

Important

Values for messages, active, dead-lettered, scheduled, completed, and abandoned messages are point-in-time values. Incoming messages that were consumed immediately after that point-in-time may not be reflected in these metrics.

Note

When a client tries to get the info about a queue or topic, the Service Bus service returns some static information like name, last updated time, created time, requires session or not etc., and some dynamic information like message counts. If the request gets throttled, the service returns the static information and empty dynamic information. That's why message counts are shown as 0 when the namespace is being throttled. This behavior is by design.

Connection metrics

Metric Name Exportable via diagnostic settings Unit Aggregation type Description Dimensions
Active Connections No Count Total The number of active connections on a namespace and on an entity in the namespace. Value for this metric is a point-in-time value. Connections that were active immediately after that point-in-time may not be reflected in the metric.
Connections Opened No Count Average The number of connections opened. Value for this metric is an aggregation, and includes all connections that were opened in the aggregation time window. Entity name
Connections Closed No Count Average The number of connections closed. Value for this metric is an aggregation, and includes all connections that were opened in the aggregation time window. Entity name

Resource usage metrics

Note

The following metrics are available only with the premium tier.

The important metrics to monitor for any outages for a premium tier namespace are: CPU usage per namespace and memory size per namespace. Set up alerts for these metrics using Azure Monitor.

The other metric you could monitor is: throttled requests. It shouldn't be an issue though as long as the namespace stays within its memory, CPU, and brokered connections limits. For more information, see Throttling in Azure Service Bus Premium tier

Metric Name Exportable via diagnostic settings Unit Aggregation type Description Dimensions
CPU usage per namespace No CPU Percent The percentage CPU usage of the namespace. Replica
Memory size usage per namespace No Memory Usage Percent The percentage memory usage of the namespace. Replica

Error metrics

Metric Name Exportable via diagnostic settings Unit Aggregation type Description Dimensions
Server Errors No Count Total The number of requests not processed because of an error in the Service Bus service over a specified period. Entity name

Operation Result
User Errors No Count Total The number of requests not processed because of user errors over a specified period. Entity name

Operation Result

Metric dimensions

Azure Service Bus supports the following dimensions for metrics in Azure Monitor. Adding dimensions to your metrics is optional. If you don't add dimensions, metrics are specified at the namespace level.

Dimension name Description
Entity Name Service Bus supports messaging entities under the namespace. With the 'Incoming Requests' metric, the Entity Name dimension will have a value of '-NamespaceOnlyMetric-' in addition to all your queues and topics. This represents the request, which was made at the namespace level. Examples include a request to list all queues/topics under the namespace or requests to entities that failed authentication or authorization.

Resource logs

This section lists the types of resource logs you can collect for Azure Service Bus.

  • Operational logs
  • Virtual network and IP filtering logs

Operational logs

Operational log entries include elements listed in the following table:

Name Description
ActivityId Internal ID, used to identify the specified activity
EventName Operation name
ResourceId Azure Resource Manager resource ID
SubscriptionId Subscription ID
EventTimeString Operation time
EventProperties Operation properties
Status Operation status
Caller Caller of operation (the Azure portal or management client)
Category OperationalLogs

Here's an example of an operational log JSON string:

{
  "ActivityId": "0000000000-0000-0000-0000-00000000000000",
  "EventName": "Create Queue",
  "resourceId": "/SUBSCRIPTIONS/<AZURE SUBSCRPTION ID>/RESOURCEGROUPS/<RESOURCE GROUP NAME>/PROVIDERS/MICROSOFT.SERVICEBUS/NAMESPACES/<SERVICE BUS NAMESPACE NAME>",
  "SubscriptionId": "0000000000-0000-0000-0000-00000000000000",
  "EventTimeString": "9/28/2016 8:40:06 PM +00:00",
  "EventProperties": "{\"SubscriptionId\":\"0000000000-0000-0000-0000-00000000000000\",\"Namespace\":\"mynamespace\",\"Via\":\"https://mynamespace.servicebus.windows.net/f8096791adb448579ee83d30e006a13e/?api-version=2016-07\",\"TrackingId\":\"5ee74c9e-72b5-4e98-97c4-08a62e56e221_G1\"}",
  "Status": "Succeeded",
  "Caller": "ServiceBus Client",
  "category": "OperationalLogs"
}

Events and operations captured in operational logs

Operational logs capture all management operations that are performed on the Azure Service Bus namespace. Data operations aren't captured, because of the high volume of data operations that are conducted on Azure Service Bus.

Note

To help you better track data operations, we recommend using client-side tracing.

The following management operations are captured in operational logs:

Scope Operation
Namespace
  • Create Namespace
  • Update Namespace
  • Delete Namespace
  • Update Namespace SharedAccess Policy
Queue
  • Create Queue
  • Update Queue
  • Delete Queue
  • AutoDelete Delete Queue
Topic
  • Create Topic
  • Update Topic
  • Delete Topic
  • AutoDelete Delete Topic
Subscription
  • Create Subscription
  • Update Subscription
  • Delete Subscription
  • AutoDelete Delete Subscription

Note

Currently, Read operations aren't tracked in the operational logs.

Virtual network and IP filtering logs

Service Bus virtual network (VNet) connection event JSON includes elements listed in the following table:

Name Description
SubscriptionId Azure subscription ID
NamespaceName Namespace name
IPAddress IP address of a client connecting to the Service Bus service
Action Action done by the Service Bus service when evaluating connection requests. Supported actions are Accept Connection and Deny Connection.
Reason Provides a reason why the action was done
Count Number of occurrences for the given action
ResourceId Azure Resource Manager resource ID.
Category ServiceBusVNetConnectionEvent

Note

Virtual network logs are generated only if the namespace allows access from selected networks or from specific IP addresses (IP filter rules).

Here's an example of a virtual network log JSON string:

{
    "SubscriptionId": "0000000-0000-0000-0000-000000000000",
    "NamespaceName": "namespace-name",
    "IPAddress": "1.2.3.4",
    "Action": "Accept Connection",
    "Reason": "IP is accepted by IPAddress filter.",
    "Count": 1,
    "ResourceId": "/SUBSCRIPTIONS/<AZURE SUBSCRPTION ID>/RESOURCEGROUPS/<RESOURCE GROUP NAME>/PROVIDERS/MICROSOFT.SERVICEBUS/NAMESPACES/<SERVICE BUS NAMESPACE NAME>",
    "Category": "ServiceBusVNetConnectionEvent"
}

Runtime audit logs

Runtime audit logs capture aggregated diagnostic information for various data plane access operations (such as send or receive messages) in Service Bus.

Note

Runtime audit logs are currently available only in the premium tier.

Runtime audit logs include the elements listed in the following table:

Name Description
ActivityId A randomly generated UUID that ensures uniqueness for the audit activity.
ActivityName Runtime operation name.
ResourceId Resource associated with the activity.
Timestamp Aggregation time.
Status Status of the activity (success or failure).
Protocol Type of the protocol associated with the operation.
AuthType Type of authentication (Azure Active Directory or SAS Policy).
AuthKey Azure Active Directory application ID or SAS policy name that's used to authenticate to a resource.
NetworkType Type of the network access: Public orPrivate.
ClientIP IP address of the client application.
Count Total number of operations performed during the aggregated period of 1 minute.
Properties Metadata that is specific to the data plane operation.
Category Log category

Here's an example of a runtime audit log entry:

{
    "ActivityId": "<activity id>",
    "ActivityName": "ConnectionOpen | Authorization | SendMessage | ReceiveMessage",
    "ResourceId": "/SUBSCRIPTIONS/xxx/RESOURCEGROUPS/<Resource Group Name>/PROVIDERS/MICROSOFT.SERVICEBUS/NAMESPACES/<Service Bus namespace>/servicebus/<service bus name>",
    "Time": "1/1/2021 8:40:06 PM +00:00",
    "Status": "Success | Failure",
    "Protocol": "AMQP | HTTP | SBMP", 
    "AuthType": "SAS | AAD", 
    "AuthId": "<AAD Application Name| SAS policy name>",
    "NetworkType": "Public | Private", 
    "ClientIp": "x.x.x.x",
    "Count": 1, 
    "Category": "RuntimeAuditLogs"
 }

Azure Monitor Logs tables

Azure Service Bus uses Kusto tables from Azure Monitor Logs. You can query these tables with Log Analytics. For a list of Kusto tables the service uses, see Azure Monitor Logs table reference.

Next steps