|
|
Sponsored Content
Special Forums
UNIX and Linux Applications
Infrastructure Monitoring
Setting Nagios warning and critical load threshold?
Post 302728365 by DGPickett on Wednesday 7th of November 2012 05:23:17 PM
11-07-2012
What do you plan to do with critical? The thresholds should be easily configurable, so you can adjust them to avoid aggravation but have them still act before they are locked out.
The load average can be high for a long time due to valid use without being a problem. For instance, an overly simple script may spawn many background jobs that punch it up until they are done, and they may be running at higher nice so other apps are still serviced. You bought the CPU with every intent of using it, and having a big queue means it never has an idle moment.
Thrashing might be more of a worry. I once wrote a simple app that read file names from stdin and mmap()'d the file for a fast string search, then unmap()'d it. I fed it with a find and rolled every other processes' pages out of RAM. Your VM may vary!
The load average can be high for a long time due to valid use without being a problem. For instance, an overly simple script may spawn many background jobs that punch it up until they are done, and they may be running at higher nice so other apps are still serviced. You bought the CPU with every intent of using it, and having a big queue means it never has an idle moment.
Thrashing might be more of a worry. I once wrote a simple app that read file names from stdin and mmap()'d the file for a fast string search, then unmap()'d it. I fed it with a find and rolled every other processes' pages out of RAM. Your VM may vary!
|
|
9 More Discussions You Might Find Interesting
1. Shell Programming and Scripting
hi all
need your help.
I am wrting a script that will load data into the table.
then on another load will append the data into the existing table.
Regards
Ankit (1 Reply)
Discussion started by: ankitgupta
1 Replies
2. UNIX for Dummies Questions & Answers
What should be the threshold for load average of a quad core processor? What constitutes "good" and "bad" load average values? (2 Replies)
Discussion started by: proactiveaditya
2 Replies
3. UNIX for Dummies Questions & Answers
Hi,
I have a table with 14 columns. How can I filter the columns 2-14, so that I get only those rows back in which the data values are >= 6 in 5 or more columns. :confused:
E.g.
A 6 6 3 6 7 8
B 1 2 3 4 5 5
C 2 2 2 6 7 8
Here I should only get back the row A.
I would like to work from... (5 Replies)
Discussion started by: danieladna
5 Replies
4. Red Hat
Hello All,
I have setup Nagios 3.2.3 on CentOS release 5.7 (Final) with the default config files and added 1 host to it and it is sending "Service Alert: CentOS 5/HTTP is WARNING" frequently, how do you fix this one? what are the additional files that need to be added so that I can monitor the... (0 Replies)
Discussion started by: lovesaikrishna
0 Replies
5. Red Hat
Hello All,
Nagios seems to be sending false alerts about few hosts, (ex: There were no users on one host and still Nagios was reporting a critical alert and says 6 users are logged in. How do I fix this one?
Also, I have installed nagios and added 12 hosts as a start and monitoring few... (4 Replies)
Discussion started by: lovesaikrishna
4 Replies
6. Solaris
Hi All
Does anybody have configured successfully nagios on a solaris 10 system?
I am having issues when running gmake all, I am having the following error:
cp: includes/rss/extlib: is a directory
*** Error code 2
make: Fatal error: Command failed for target `install'
Current working... (0 Replies)
Discussion started by: fretagi
0 Replies
7. Infrastructure Monitoring
We have a dual Nagios server setup. One is setup for internal server monitoring on our LAN, while the second Nagios server is hosted externally and is used for external checks only such as URL and ping checks form the WAN side.
I was wondering if there is any way to setup cross dependencies... (1 Reply)
Discussion started by: eugenes18t
1 Replies
8. UNIX for Dummies Questions & Answers
Hello, on my hostserver i see one VPS of mine got load of 200.00 and netstat nothing (not a single blank line on netstat command) after some time, netstat started showing connections, but i see no excessive IP connections.
tail -f /var/log/httpd/access_log shows no activity
/var/log/messages ;... (1 Reply)
Discussion started by: postcd
1 Replies
9. AIX
This's my problem
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
LC_ALL = "en_US.UTF-8",
LC__FASTMSG = "true",
LC_MESSAGES = "",
LC_CTYPE = "en_US.UTF-8",
LC_TYPE = "en_US.UTF-8",
LANG = "EN_US"... (1 Reply)
Discussion started by: bobochacha29
1 Replies
LEARN ABOUT HPUX
ruptime
ruptime(1) General Commands Manual ruptime(1) NAME
ruptime - show status of local machines SYNOPSIS
DESCRIPTION
outputs a status line for each machine on the local network that is running the daemon. status lines are formed from packets broadcast once every 3 minutes between daemons (see rwhod(1M)) on each host on the network. Each status line has a field for the name of the machine, the status of the machine (up or down), how long the machine has been up or down, the number of users logged into the machine, and the 1-, 5- and 15-minute load averages for the machine when the packet was sent. The status of the machine is reported as ``up'' unless no report has been received from the machine for 11 minutes or more. The length of time that the machine has been up is shown as: Load averages are the average number of jobs in the run queue over the last 1-, 5- and 15-minute intervals when the packet was sent. An example status line output by might be: The above status line would be interpreted as follows: is presently ``up'' and has been up for 1 day, 5 hours and 15 minutes. It currently has 7 users logged in. Over the last 1-minute inter- val, an average of 1.47 jobs were in the run queue. Over the last 5-minute interval, an average of 1.16 jobs were in the run queue. Over the last 15-minute interval, an average of 0.80 jobs were in the run queue. If a user has not used the system for an hour or more, the user is considered idle. Idle users are not shown unless the option is speci- fied. Options If no options are specified, the listing is sorted by host name. Options change sorting order as follows: Sort by load average. Sort by up time. Sort by the number of users. Reverse the sort order. DIAGNOSTICS
No status report files in Ask the system administrator to check whether the daemon is running. AUTHOR
was developed by the University of California, Berkeley. FILES
Data files SEE ALSO
rwho(1), rwhod(1M). ruptime(1)