Now accepting guest editorial from SAP, Oracle, Workday and ServiceNow practitioners.Pitch a guide
The Sign-Off Table

Lesson 6 of 6

Monitoring and diagnosis

9 min

Listen to this lesson

0:00 / 7:10

Monitoring is how you prove an integration works and how you find out quickly when it stops. In the Integration Suite home page, the Monitor area shows message processing per integration flow, the state of deployed content, security material, stores such as data stores and message queues, and logs. Learning to read it well is the fastest way to become the person teams call.

Message statuses

StatusWhat it meansWhat to do
CompletedThe flow finished without an unhandled errorFor tests, still check the business result at the receiver
ProcessingThe message is still runningNormal for long flows; investigate if it never moves on
RetryProcessing failed and a retry is pendingCheck the error; watch whether retries will succeed
FailedProcessing ended with an errorOpen the log, find the failing step, classify the error
EscalatedProcessing ended through an escalation path in error handlingTreat as a business-relevant failure and follow the flow's design
CancelledProcessing was stopped, for example manuallyConfirm who cancelled it and whether the message needs reprocessing

Message statuses you will see in Monitor Message Processing. Check your tenant's documentation for the full list.

Message statuses in Monitor Message Processing: Retry loops back, while Completed, Failed, Escalated and Cancelled are end states that each ask for a different action.
The Sign-Off Table · thesignofftable.com/academy/sap-int-203

Reading the message processing log

Each message has a message processing log (MPL): start and end time, status, the steps it passed through, the error text, and custom header values such as the business key you set. The amount of step detail depends on the flow's log level. SAP Learning material lists None, Info (the default), Debug and Trace. Trace also records message content at each step, but only for a limited period: SAP describes trace as switching off after about ten minutes, with the captured content kept for a short time, and access to payloads restricted to users with the right role.

  1. Find the messageFilter by flow and time, or search by the business key header, such as PO 4500002210.
  2. Read the status and errorThe log names the failing step and the error text, for example an authentication failure at the SFTP receiver.
  3. ClassifyTransient, data, configuration or unknown, as in lesson 4.
  4. Reproduce with trace if neededSet Trace briefly, resend a test message, inspect the payload before and after each step.
  5. Fix and proveCorrect the cause, reprocess or resend, and confirm Completed plus the business result.
A diagnosis routine for a failed message in Cloud Integration.
The Sign-Off Table · thesignofftable.com/academy/sap-int-203

Trace is a diagnostic tool, not a setting to leave on. It records payloads, which may contain personal and commercial data, and it consumes tenant resources. Turn it on for a specific problem, capture what you need, and let it expire.

Beyond one message

Message-by-message monitoring does not scale to about 60 interfaces. For the program view, SAP Cloud ALM's Integration and Exception Monitoring collects messages from SAP Integration Suite and other components, correlates them into end-to-end flows, lets users search by business attributes such as an order number, and raises alerts through rules that can email, notify, create a ticket or start an operation flow. Contoso's S/4HANA program already runs in SAP Cloud ALM, so that is where Theo sets up interface alerting for hypercare.

SAP Cloud ALM turns a failed message into an alert for a named owner, who then works from the business key to the message log.
The Sign-Off Table · thesignofftable.com/academy/sap-int-203
SAP Cloud ALMIntegration and Exception Monitoring: end-to-end correlation, alert rules, assignment, trends
Integration Suite monitorMessage processing logs, deployed content, stores, security material, connectivity tests
S/4HANA monitorsIDoc status, queues, web service logs on CMP
Partner sideLakeside's receipt, acknowledgments and rejections
Layers of monitoring for one Contoso interface, from program view down to the partner.
The Sign-Off Table · thesignofftable.com/academy/sap-int-203

What would you do?

A Lakeside order message fails with an authentication error at the SFTP receiver in the production tenant. It worked yesterday. Luis asks Theo for a quick answer. What should Theo check first?

Key points
  • Know the message statuses and what each one asks you to do.
  • The message processing log names the failing step; set a business key header so you can search.
  • Use Trace briefly for diagnosis; it records payloads and expires.
  • Use SAP Cloud ALM Integration and Exception Monitoring for end-to-end views and alert rules.

Check your understanding · not graded

Lesson check

  1. 1. Which log level records message content at each step for a limited time?

  2. 2. A message shows Retry. What does that tell you?

  3. 3. Why set a business key, such as the PO number, as a header in the flow?

  4. 4. Which SAP tool correlates messages into end-to-end flows and raises alerts by rule?