Instrumenting the consumer
While you might be able to get away without custom instrumentation on the producer, consumer instrumentation is unavoidable.
Some brokers push messages to consumers using synchronous HTTP or RPC calls, and the existing framework instrumentation can provide the bare minimum of observability data. In all other cases, messaging traces and metrics are all we have to detect consumer health and debug issues.
Let’s start by tracing individual messages – recording when they arrive in the consumer and how they are processed. This allows us to debug issues by answering questions such as “Where is this message now?” or “Why did it take so long to process the data?”
Tracing consumer operations
When using Azure Queue Storage, applications request one or more messages from the queue. Received messages stay in the queue but become invisible to other consumers for configurable visibility timeout. The application processes...