# Usage

The following metrics are published to Application Insights

The name of the metric is EventHubPartitionLag. The value of the metric is the number of messages that have not yet been processed in the context of a certain event hub, partition and consumer group.

Custom dimensions of this metric are:

• EventHubNamespace: Name of the Event Hub Namespace.
• EventHub: Name of the Event Hub.
• ConsumerGroup: Name of the consumer group.
• PartitionId: Numerical identifier of the partition. If you have 4 partitions, values are 0, 1, 2 and 3.

This metric provides maximum flexibility as it includes information about partitions.

The lag for a particular Event Hub and Consumer Group can be extracted using the following Log Analytics / KQL query:

customMetrics
| where name == 'EventHubPartitionLag'
| extend eventHub=tostring(customDimensions['EventHub'])
| extend consumerGroup=tostring(customDimensions['ConsumerGroup'])
| extend partitionId=tostring(customDimensions['PartitionId'])
| where consumerGroup == '$Default' and eventHub == 'example-event-hub'
| summarize lag=sum(value) by bin(timestamp, 1m)
| order by timestamp desc

The name of the metric is EventHubLag. The value of the metric is the number of messages that have not yet been processed in the context of a certain event hub and consumer group. The values are already summed up over all partitions of the Event Hub. This level of detail is usually sufficient, as the overall lag across all partitions is most relevant.

Custom dimensions of this metric are:

• EventHubNamespace: Name of the Event Hub Namespace.
• EventHub: Name of the Event Hub.
• ConsumerGroup: Name of the consumer group.