Job and automation monitoring

Much of an SAP system's work happens when nobody is looking: MRP runs, billing runs, payment runs, interface jobs, conversion follow-up jobs. A batch chain that fails at 02:00 and is noticed at 09:00 costs a business day. During a cutover it can cost the go-live window. Job & Automation Monitoring gives a central view of job executions across SAP products, so the people responsible see failures and delays as they happen.

A monitor, not a scheduler

SAP describes Job & Automation Monitoring as a central job monitoring solution that complements local job scheduling. It does not schedule anything. Jobs are still scheduled where they run, in SM36 or the Application Jobs app on S/4HANA, or in another scheduler. Cloud ALM collects execution data and rates it. SAP's content list covers ABAP jobs, application jobs, BW process chains, SAP Build process automation jobs, workflow, BTP job scheduling and more.

Two statuses and two KPIs

Each execution is rated on two statuses. Execution status says whether the run completed: green for success, yellow for finished with warnings or errors without aborting, red for aborted. Application status says whether the application data was processed correctly, based on messages such as ABAP application log errors. A job can finish green and still be red on application status: it ran to the end but failed to post half its documents.

Two KPIs complete the picture. Start delay is the gap between planned and actual start. Run time is rated against a threshold. In a cutover, both matter: a job that starts 40 minutes late because its predecessor overran pushes every later step.

Batch tests and cutover

Contoso's conversion needs batch evidence twice: in SIT, where the nightly chains run against converted data in CMQ, and in the two mock conversions and the dress rehearsal, where Luis times every cutover step against the 60-hour window from Friday 18:00 to Monday 06:00.

For each mock conversion, Luis records the actual runtimes from the monitor into the cutover plan. By the dress rehearsal, the runtime thresholds in Cloud ALM are set from real data, so an overrun during the live cutover raises an alert instead of surprising someone at the go/no-go meeting.

Think in chains, not single jobs. Contoso's nightly order-to-cash chain runs delivery creation, then billing, then the EDI invoice output, then the interface job that sends it. A failure early in the chain often shows up later as a missing message or an empty billing run. When a job alert fires, the integration specialist checks the downstream scenarios in Integration & Exception Monitoring before declaring the problem fixed, and Luis's cutover plan names the predecessor and successor of every monitored job so the team knows which later steps to hold.

Data retention

Keep evidence before it ages out. SAP's setup guide states that individual executions are kept for two days by default (configurable from one to seven days, with at least the last 15 executions per job kept), while hourly aggregates are kept much longer. If the cutover report needs exact runtimes, export or record them during the weekend.