A teal-inlet fleet streams click telemetry that several analytics jobs must read independently. Which service should the developer pick?
Select an answer to reveal the explanation.
Short Explanation
Many readers, one firehose—that’s Event Hubs. Don’t jam click telemetry into a Service Bus queue or stretch Event Grid into a stream.
Full Explanation
When several analytics jobs must read the same teal-inlet click telemetry independently, Event Hubs is the append-only high-throughput stream with consumer groups that each keep their own offsets. Multiple readers progress through the retained stream without deleting events out from under each other, which matches fleet-scale click telemetry better than competing queues or discrete reaction topics.
A Service Bus queue that deletes each message after one consumer fails independent fan-out of a stream: queues implement competing consumers, so one receiver completes a message and the others never see that same click for separate analytics.
Event Grid only, treating every click as a one-off discrete reaction, fails the streaming pattern: high-volume click telemetry is a continuous firehose, not a sparse “something happened” reaction topic that Event Grid is optimized to route.
Blob lease locks polled by each job as the streaming fabric fails because leases coordinate exclusive access to blobs; they are not Event Hubs partitions and consumer groups for streaming telemetry ingestion and replay.
Exam caveat: Event Hubs retains events for a configured time window—late joiners can rewind within retention, unlike a queue that drops messages after settle. Operational check: create two consumer groups, run two jobs against the same hub, and confirm each job’s checkpoint advances independently while both see the full click stream within retention.