BTP & Integrationintermediate

BTP Event Mesh Part 3: Reliable Event Consumption

Key takeaway

Consume events without losing them: persistent queues, acknowledgements after successful processing, redelivery on failure, maximum redelivery and dead message queues, consumer scaling, and monitoring queue depth.

Explanation

📌 KEY POINTS
✅ Events wait in queues until consumers acknowledge them.
✅ Acknowledge only after successful processing; failures lead to redelivery.
✅ After max redeliveries, messages move to a dead message queue for analysis.
✅ Monitor queue depth; scale consumers when backlog grows.
⚠️ Acknowledging before processing (auto-ack) loses events when the consumer crashes mid-processing.

━━━━━━━━━━━━━━━━━━━━
▶ WHERE WE ARE
Part 2 configured Event Mesh. Now reliability.

━━━━━━━━━━━━━━━━━━━━
▶ 1. PERSISTENCE

Queues store events durably.

━━━━━━━━━━━━━━━━━━━━
▶ 2. ACKNOWLEDGEMENT

Manual ack after success.

━━━━━━━━━━━━━━━━━━━━
▶ 3. REDELIVERY

Retry behavior; limits.

━━━━━━━━━━━━━━━━━━━━
▶ 4. DEAD MESSAGES

DMQ for poison messages.

━━━━━━━━━━━━━━━━━━━━
▶ 5. MONITORING & SCALING

Queue depth, consumer count.

━━━━━━━━━━━━━━━━━━━━
▶ STEP BY STEP

Consume with manual ack.
Simulate a failure and observe redelivery.
Configure max redelivery and DMQ.

✅ CHECKPOINT: poison event lands in DMQ; others continue.

Code example

ABAP Code

===== CONSUMER LOGIC (pseudo) =====receive(msg)try { process(msg); ack(msg) }catch (e) { nack(msg) }   // redelivered until max redelivery → dead message queue ===== QUEUE SETTINGS (examples) =====maxRedeliveryCount: 5 | deadMsgQueue: acme/orders/v1/crm-updates-dmq | message expiry as needed

Real project scenario

A consumer auto-acknowledged events and then crashed during S/4 updates; several customer changes were lost. Switching to ack-after-success and adding a DMQ ensured no event was lost.

Common mistakes

Auto-ack.
No DMQ.
Unlimited redelivery.
No depth monitoring.
Single consumer bottleneck.

Best practices

Ack after success.
DMQ.
Redelivery limits.
Depth alerts.
Scale consumers.

Interview angle

Questions: "When should a consumer acknowledge?", "What is a dead message queue?", "How do you detect consumer problems?"
Strong answer: persistence, ack, redelivery, DMQ, monitoring.

Practice this lesson

Quick check

1. What typically happens to a message on an Event Mesh queue after it repeatedly fails delivery to consumers beyond the configured retry limit?

2. A consumer auto-acknowledged events and then crashed; changes were lost. What fixes this?

3. In SAP BTP Event Mesh, what is the key functional difference between a queue and a topic?

Interview questions

1. A production issue is reported in Event Mesh. What would you check first, and how would you troubleshoot it systematically?

Answer it in your own words first, then compare.

2. Events are occasionally lost or block processing. How do you make consumption reliable?