Black Box 1102, Secure Device Servers, 1101 Basic Nagios Plug-Ins, Number of Supported Devices

Chapter 10: Nagios Integration

use

check_command active_checks_enabled passive_checks_enabled

}

generic-service

check_conn_via_Black Box!tcp!22 0 1

define servicedependency{
name	Black Box_nrpe_daemon_dep
host_name	Black Box
dependent_host_name	server
dependent_service_description	SSH Port

service_description	NRPE Daemon
execution_failure_criteria	w,u,c
}

10.4.2 Basic Nagios Plug-Ins

Plug-ins are compiled executables or scripts that can be scheduled to run on the console server to check the status of a connected host or service. This status is then communicated to the upstream Nagios server that uses the results to monitor the current status of the distributed network. Each console server is preconfigured with a selection of the checks that are part of the Nagios plug-ins package:

check_tcp and check_udp are used to check open ports on network hosts

check_ping is used to check network host availability

check_nrpe is used to execute arbitrary plug-ins in other devices

Each console server is preconfigured with two checks that are specific to Black Box:

check_serial_signals is used to monitor the handshaking lines on the serial ports

check_port_log is used to monitor the data logged for a serial port.

10.4.3 Number of Supported Devices

Ultimately the number of devices any particular console server can support depends upon the number of checks made, and how often they are performed. Access method will also play a part. The table below shows the performance of three of the console servers:

Time		No encryption		3DES		SSH tunnel
NSCA for single check		~ ½ second		~ ½ second		~ ½ second
NSCA for 100 sequential checks		100 seconds		100 seconds		100 seconds
NSCA for 10 sequential checks, batched upload		1 ½ seconds		2 seconds		1 second
NSCA for 100 sequential checks, batched upload		7 seconds		11 seconds		6 seconds
	No encryption		SSL		no encryption -
					tunneled over existing
					SSH session
NRPE time to service 1 check	1/10th second		1/3rd second		1/8th second
NRPE time to service 10 simultaneous	1 second		3 seconds		1 ¼ seconds
checks
Maximum number of simultaneous checks	30		20 (1,2 and 8) or 25		25 (8 port), 35 (16 and
before timeouts			(16 and 48 port)		48 port)

The results were from running tests 5 times in succession with no timeouts on any runs. There are a number of ways to increase the number of checks you can do.

Usually when using NRPE checks, an individual request will need to set up and tear down an SSL connection. This overhead can be avoided by setting up an SSH session to the console server and tunneling the NRPE port. This allows the NRPE daemon to run securely without SSL encryption, because SSH will provide the security.

When the console server submits NSCA results, it staggers them over a certain time period (for example, 20 checks over 10 minutes will result in two check results every minute). Staggering the results like this means that if the power fails or other incident causes multiple problems, the individual freshness checks will be staggered too.

NSCA checks are also batched. In the previous example, the two checks per minute are sent through in a single transaction.

724-746-5500 blackbox.com 103