Increase checkmk check interval to a longer interval in order to save CPU resources on the server and target systems

Try to lower the frequency of checkmk runs.

• Search check interval from the "Setup" from checkmk interface

• Create a rule for "Normal check interval for service checks"

• Create a rule for "Normal check interval for host checks

References:

https://docs.checkmk.com/latest/en/monitoring_basics.html
https://forum.checkmk.com/t/solved-change-check-interval-of-my-own-datasource-program-checks/20312
https://forum.checkmk.com/t/check-mk-english-how-to-set-frequency-of-mk-apt-plugin/12123

Suggestion: disable irqbalance service on Virtual Guests (Virtual Machine)

irqbalance is a Linux daemon that distributes interrupts over among the processors and cores in your computer system. The design goal of irqbalance is to do find a balance between power savings and optimal performance. To a large degree, the work irqbalance does is invisible to you. The daemon balances savings in power consumption with performance. On Virtual Guests (Virtual Machine). It does not really make sense because unless you are pinning the guest to specific CPUs and IRQs and dedicated net/storage hardware, you will likely not see the benefits you would on bare metal. On my Debian VM:

mds@debian:~# dpkg -l|grep irq
ii  irqbalance                    1.5.0-3                      amd64        Daemon to balance interrupts for SMP systems
root@download:~# dpkg --purge irqbalance
(Reading database ... 32086 files and directories currently installed.)
Removing irqbalance (1.5.0-3) ...
Purging configuration files for irqbalance (1.5.0-3) ...
Processing triggers for systemd (241-7~deb10u7) ...

Delete trillion records from a large table

Try to clean a large table (non partition, over 5 trillion records) on a old Oracle 10g Database. [TYPE_ID] in the PK. First try:

DELETE FROM LARGE_TABLE_T
WHERE TYPE_ID = 199
AND ROWNUM < 1000;

COMMIT;

999 rows deleted. Elapsed: 00:00:11.42		90/second

DELETE FROM LARGE_TABLE_T
WHERE TYPE_ID = 199
AND ROWNUM < 10000;

COMMIT;

9999 rows deleted. Elapsed: 00:00:19.54		526/second

DELETE FROM LARGE_TABLE_T
WHERE TYPE_ID = 199
AND ROWNUM < 100000;

COMMIT;

99999 rows deleted. Elapsed: 00:04:47.26	348/second


DELETE FROM LARGE_TABLE_T
WHERE TYPE_ID = 199
AND ROWNUM < 20000;

COMMIT;

19999 rows deleted. Elapsed: 00:00:40.08	500/second

DELETE FROM LARGE_TABLE_T
WHERE TYPE_ID = 199
AND ROWNUM < 5000;

COMMIT;

4999 rows deleted. Elapsed: 00:00:10.04		499/second


DELETE FROM LARGE_TABLE_T
WHERE TYPE_ID = 199
AND ROWNUM < 9000;

COMMIT;

8999 rows deleted. Elapsed: 00:00:21.52		409/second

DELETE FROM LARGE_TABLE_T
WHERE TYPE_ID = 199
AND ROWNUM < 50000;

COMMIT;

49999 rows deleted. Elapsed: 00:01:51.70	447/second

I wrote a quick loop with commit every 10000 rows:

BEGIN
LOOP 
  DELETE FROM LARGE_TABLE_T WHERE TYPE_ID = 199 AND ROWNUM < 10000;
  COMMIT;
END LOOP;
COMMIT;
END;

Not bad, Eh?
I let it run on the sqlplus from server and keep try different way.

Found another way faster solution after few tweaks:

DECLARE

  CURSOR c_delete_rowids
  IS
	SELECT ROWID
    FROM LARGE_TABLE_T
	WHERE TYPE_ID = 199
	AND ROWNUM <= 1000000;

  v_count number(4) default 0;

BEGIN

  FOR rec IN c_delete_rowids
  LOOP
  
    delete from LARGE_TABLE_T where rowid = rec.rowid;
	v_count := v_count + 1;

	IF v_count >= 1000 then
		v_count := 0;
		COMMIT;
	END IF;

  END LOOP;

  COMMIT;

END;
/

Which delete 1,000,000 records in 1 minute.

SQL> 13:26:21 SQL> 13:26:21 SQL> 13:26:21 SQL> 13:26:21   2  13:26:21   3  13:26:21   4  13:26:21   5  13:26:21   6  13:26:21   7  13:26:21   8  13:26:21   9  13:26:21  10  13:26:21  11  13:26:21  12  13:26:21  13  13:26:21  14  13:26:21  15  13:26:21  16  13:26:21  17  13:26:21  18  13:26:21  19  13:26:21  20  13:26:21  21  13:26:21  22  13:26:21  23  13:26:21  24  13:26:21  25  13:26:21  26  13:26:21  27  13:26:21  28  13:26:21  29  13:26:21  30  13:26:21  31
PL/SQL procedure successfully completed.

Elapsed: 00:01:02.75

To protect the archive log not run crazy, I call it from a shell script and sleep 100 every loop. 1.5 trillion records has been delete around 3 days. Super!!!