System Monitoring & Troubleshooting

ST22 shows a spike of dumps. How does the pattern many dumps versus dumps from a single user change where you look?

You need a deeper message-server trace during a live connectivity issue. What can SMMS change without a restart, and why does that matter?

Table SNAP is growing without bound and you're mid-investigation on one critical dump. How do you stop the growth without losing that dump?

You suspect a user is logged on somewhere holding a lock. SM04 versus AL08 which one actually answers that, and why can SM04 mislead?

In SM50 a work process shows internal number 8, PID 4711, and type DIA. To open its developer trace, which number do you use and why?

You need to move a trace file off the app server to your desktop, and separately push a cofile up to the server. Which is CG3Y and which is CG3Z?

SM51 lists all your application servers but one shows a non-Active status. Why does that matter, and what does SM51 actually tell you per server?

A work process crashed, you restarted the instance, and now you want the trace from the run before the restart. Is it still there?

The GUI is unreachable but you need to inspect the gateway from the OS. Which command, and what's its syntax?

Nobody scheduled a job to warn about expiring certificates, yet warnings still appear. What produces them, and why can't you manage them like a normal SM37 job?

An update work process appears stuck. Why is manually cancelling it the wrong instinct, and what's the right way if you must?

A security incident needs a record of a user's actions. Why is the SM04 user trace the wrong source, and what is the right one?

The local application-server syslog file has hit its size limit. What happens to new log entries, and what does that imply for forensics?

A work process sits in On Hold reason PRIV and won't free up. What does PRIV actually mean, and how does it differ from SLEEP and VB?

Name the three OS-level programs that monitor the message server and what each does.

You raised rdisp/TRACE to 2 to chase an intermittent error and forgot to lower it. What's the cost, and what should the steady-state value be?

A dialog step is slow. ST06 versus an ABAP-level trace which question does ST06 actually answer?

The system was hanging and the kernel created a snapshot on its own. What's in it, and why is that a lucky break for diagnosis?

SM21 is your first stop for an incident. Besides red, which SM21 color should you NOT ignore, and why?

In SM50, which work-process priority maps to dialog, which to RFC, and which to background and what's the common backwards assumption?

An ABAP program just terminated on an unhandled runtime exception. What got captured, where does it live, and how should you actually read it?

You have no OS access but need to run an OS command from SAP GUI. What report do you use, and what must you never run through it?

A user hits a 'no authorization' error. ST01 versus STAUTHTRACE which do you turn on, and what's the cost of picking ST01?

You want to query the message server from a browser. What's the configuration step people skip, and where do you configure it?

Your OS performance data in SAP suddenly goes stale or empty. What single component feeds it, and what does that tell you to check first?

Every dialog work process on an instance is occupied so you can't even log on to that instance. What OS tool gets you the work process table, and what can it do that the GUI now can't?

You must drop a misbehaving external RFC connection via SMGW. What does SMGW administer, and when is soft versus hard shutdown the right call?

You want to drain an application server for maintenance without dropping work in progress. What does SM51 Deactivate do, and what does it NOT do?

SM50 only shows the local instance. To hunt a runaway process across all instances, which transaction do you use, and what filter makes it usable?

Several component traces exist (dev_disp, dev_icm, dev_ms, dev_rd, dev_tp). A transport import is failing at the OS level which one do you open, and how do you read them in the GUI?

Users complain a transaction is 'slow' but the ABAP work-process times look fine; the delay might be the network or the browser. Which Cloud ALM capability pins the blame on the right component?

Your team only catches failed outbound emails by manually opening SOST every couple of days. What modern option lets an S/4HANA admin be alerted to send problems proactively?

In an S/4HANA SaaS operating model there is no classic interface-monitoring transaction to fall back on. How does SAP fill that gap, and what limits the free version?

Your shop still equates 'SAP monitoring' with Solution Manager. In a hybrid S/4HANA landscape, which product does SAP now push as strategic for monitoring and operating the system?

You want an email the moment short dumps spike on your S/4HANA system. Walking through Cloud ALM Health Monitoring, what do you actually have to set for that alert to fire?

After a partner interface silently stalled, you want automatic alerts when an RFC destination fails, not just eyeballing SM58/SMQ2. How does an S/4HANA admin set that up through Cloud ALM, and what is the catch?

The Web Dispatcher needs an external load balancer for redundancy, but what does the Cloud Connector offer for high availability, and where does its admin see certificate or connection alarms?

Before reaching for Cloud ALM, S/4HANA ships an on-stack cockpit that watches host, database and AS ABAP from one screen. What is it, and how do you open its web version?

There is no 24/7 shift, and users report the system crawled during last night's batch window. Which tool lets you reconstruct what the work processes were doing after the fact?

You configured HANA alert thresholds in the DBA Cockpit, but the alerts monitored through Cloud ALM never change. Why, and where should those thresholds live?

A HANA incident spikes CPU at 3 a.m. and is gone by morning with no diagnostics captured. Which tool can auto-collect dumps when thresholds are crossed, and what is the licensing caveat?

The built-in 10x10 MB trace rotation still leaves explicitly created SQL traces and RTE dumps filling the file system. Which tool automates cleaning those up?

Beyond checking SM37 yourself, how can an S/4HANA admin be told automatically when a critical background job fails or overruns?

You post an SM02 system message about downtime in two hours. A user logs on, sees it, logs off, logs back on an hour later — do they see it again?

You fixed the condition behind a SAP HANA alert but it is still showing. Why, and how long will old alerts linger anyway?