How to monitor?
First you have to know what to monitor. To be honest, I don't know.
Anyway, just blindly monitor:
Article Directory
SM50/SM66 watch system process
View the application server:
Open on this one instance, double-click on the go will be able to see the progress of this service instance:
is SM50 , then you will see all the processes on the server this we. The ones that are running and waiting are usually only these two states. If there are other states, there is a problem. There are 77 processes in total, which are occupied and which are not. Can be seen.
There is an action in the last column, you can see what the process has done:
if the DIA process is full, you may not be able to log in and pull, or you may be slow to operate. Because you have to wait for the process. Or there are processes that have been running for a long time, and they are all red under the time column. If there is this, you have to kill it manually: select this process: cancel->without core. Cancel the job directly, and release the process to others.
If there is a PRIV state, it may be that the operation takes up too much memory and other DIAs can't run. In the end, stop DIA and run in the background.
If there is an SPO print job that takes a long time, then see if there is a problem with the print server or printer.
SM50 looks at a single system process. And can come to kill the process.
Then SM66 looks at the process status across systems. If it can't break, then you can use this system. . . We don’t know how to get it. . .
AL08/SM04 View user data
al08 is equivalent to sm66, cross-system.
sm04 depends on the current system. I can see who, what I did, and how to connect.
The RFC estimate is in the report, and the GUI is operating in the system.
SM12/SM13 View lock table and system update
Looking at the lock with sm12, this lock mechanism means that while you are doing it, others cannot change it at the same time. After you change it, the lock is released. Then they can change it.
If you change the network and disconnect in the middle, then your lock will not be released. It may be impossible for others to access or modify it. This only keeps the lock table up to the previous day's timestamp.
Specific lock mechanism: lock mechanism
SM13 to see the system update, if there is no update, it is empty, if there is an update, then check whether the update is normal. Look through the administration.
SM21/ST02/ST06 system log monitoring, tuning, CPU utilization
SM21 to check the system log:
see what happened:
ST02 to see the tuning summary:
if you see red under SWAPS:
then double-click to find the current parameter:
then look at the current parameter value to
directly change the parameter value, Or RZ10 to change, and then restart the system. Only restart the parameter value will take effect.
From ST06, the CPU utilization
idle idle CPU cannot be lower than 20%.
Under normal circumstances, the CPU utilization is too high, you have to go to SM50 to check whether there is always a job occupying memory.
Then you can check the lock on SM12, check whether there is an update on SM13, and check the system log on SM21. You can watch it anyway.
ST22 system dump
I won’t talk about this, just go to see what system dump was yesterday or today, and then double-click to see what it is and find out what solutions are there.
SM37 job monitoring
DB02 database monitoring
DB12 backup log
Check whether the backup is successful, and select it after entering DB02.
Others:
SP01: spool request monitoring
SM58: tRFC monitoring
Summary:
DB12 checks the database backup.
Is SM51 enabled on all servers?
SM50 checks whether the workflow is in both waiting and running status.
SM66 global workflow, to see if there is a long time running, like the one that exceeds 1800 seconds.
SM13 View and update the
SM21 system log for errors, warnings, security, and database issues.
SM37 Check whether there is an error in the job.
SM12 See if there is a lock that has not been released for a long time.
SM04/AL08 View what the system user is doing.
DB02 Check the table space. If the table space usage exceeds 90%, it is best to rebuild the index or something.