Introduction
- System Monitors are tools that execute repeated checks for availability of software components.
- For example, Nagios® is an Open Source System Monitor that is available at http://www.nagios.org.
- An ecosystem of more or less compatible System Monitors has evolved, e.g. op5®, Checkmk and many others.
- Users can monitor JS7 by use of Open Source licensed products and by commercially licensed products that check the availability of the JOC Cockpit, Controller and Agents..
Active Checks and Passive Checks
- Active Checks are initiated from the System Monitor server and are performed on a regular basis, e.g. every 5 minutes.
- The only purpose of Active Checks is to verify the availability of the JobScheduler application.
- This type of check does not provide information about the execution status of individual jobs.
- The System Monitor Plugins explained in this article are intended for Active Checks only.
- Passive Checks are initiated from the JOC Cockpit and report individual execution results to a System Monitor.
- The JobScheduler Monitoring Interface is a complete solution for the management of complex Passive Checks.
Download and Installation of the Plugins
Download
The following Plugins are available for download:
- JS7 Controller: https://download.sos-berlin.com/SystemMonitor/check_js7_controller.pl
- JS7 Agents:: https://download.sos-berlin.com/SystemMonitor/check_js7_agent.pl
Both Plugins include to check availability of JOC Cockpit.
Installation
Copy the files to the plugin directory of your System Monitor installation.
Configuration of the Plugins
The configuration for the JS7 Controller and Agent Plugins includes to define
- a System Monitor Command that specifies the parameterization of the Plugin and
- a System Monitor Service that is the visible point of control in the System Monitor console.
JobScheduler Controller Plugin
Sample Configuration
# 'check_jobscheduler_with_joc' command definition define command { command_name check_js7_controller command_line /opt/plugins/check_js7_controller.pl -j $ARG1$ -i $ARG2$ -a $ARG3$ -t $ARG4$ }
define service { use generic-service host_name localhost service_description JS7Controller is_volatile 0 check_period 24x7 max_check_attempts 1 normal_check_interval 1 retry_check_interval 1 contact_groups admins notification_options w,u,c,r notification_interval 960 notification_period 24x7 check_command check_js7_controller!http://localhost:4446!jobscheduler!test:test!5! active_checks_enabled 1 passive_checks_enabled 0 }
Parameterization
Parameter | Data Type | Default | Description | |
---|---|---|---|---|
-j | --joc-url | String | — | URL of JOC Cockpit (HTTP and HTTPS are supported) |
-i | --id | String | — | ID of a JS7 Controller |
-a | --account | String | — | Account for HTTP authentication with JOC Cockpit (<user:password>) |
-c | --cafile | String | — | Location of a certificate file or a CA bundle file to verify the SSL connection to JOC Cockpit |
-C | --capath | String | — | Path to a directory that includes certificate files or CA bundle files to verify the SSL connection to JOC Cockpit |
-s | --hostname-verification | n/a | false | This switch activates SSL hostname verification |
-t | --timeout | Number | 30 | Timeout in seconds for establishing the connection to JOC Cockpit |
-d | --detailed | n/a | false | If set then status information for each Controller instance is returned. A Cluster can include two Controller instances. |
Example:
# The JS7 Controller is operated for the ID 'jobscheduler' and JOC Cockpit has the URL http://localhost:4446 with the account (user and password 'test') enabled <path_to_plugins>/check_js7_controller.pl -j http://localhost:4446 -i jobscheduler -a test:test
Notes
- JOC Cockpit installation since 2.0.0 is required.
Perl > 5.8 is required and some Perl packages are required. You can download and install Perl packages from e.g. http://www.cpan.
HTTP::Request
LWP::UserAgent
LWP::Protocol::https
(if HTTPS is used to connect to JOC Cockpit)JSON
MIME::Base64
JobScheduler Agent Plugin
A service Command has to be declared before configuring the System Monitor Service that makes use of this Command.
Sample Configuration
# 'check_js7_agent' command definition define command { command_name check_js7_agent command_line /opt/plugins/check_js7_agent.pl -j $ARG1$ -i $ARG2$ -a $ARG3$ -t $ARG4$ }
define service { use generic-service host_name localhost service_description JS7Agents is_volatile 0 check_period 24x7 max_check_attempts 1 normal_check_interval 1 retry_check_interval 1 contact_groups admins notification_options w,u,c,r notification_interval 960 notification_period 24x7 check_command check_js7_agent!http://localhost:4446!jobscheduler!test:test!5! active_checks_enabled 1 passive_checks_enabled 0 }
Parameterization
Parameter | Data Type | Default | Description | |
---|---|---|---|---|
-j | --joc-url | String | — | URL of JOC Cockpit (HTTP and HTTPS are supported) |
-i | --id | String | — | ID of a JS7 Controller |
-a | --account | String | — | Account for HTTP authentication to JOC Cockpit (<user:password>) |
-A | --agent | — | URL of an Agent, optional, can be specified several times | |
-c | --cafile | String | — | Location of a certificate file or a CA bundle file to verify the SSL connection to JOC Cockpit |
-C | --capath | String | — | Path to a directory that includes certificate files or CA bundle files to verify the SSL connection to JOC Cockpit |
-s | --hostname-verification | n/a | false | This switch activates SSL hostname verification |
-t | --timeout | Number | 30 | Timeout in seconds for establishing the connection to JOC Cockpit |
-d | --detailed | n/a | false | If set and the Cluster has more than one Master and not all Masters are running then the message contains host:port of each Master |
Example:
# The JS7 Controller is operated for the ID 'jobscheduler' and JOC Cockpit has the URL http://localhost:4446 with the account (user and password 'test') enabled <path_to_plugins>/check_js7_agent.pl -j http://localhost:4446 -i jobscheduler -a test:test
Notes
- JOC Cockpit installation since 2.0.0 is required.
Perl > 5.8 is required and some Perl packages are required. You can download and install Perl packages from e.g. http://www.cpan.
HTTP::Request
LWP::UserAgent
LWP::Protocol::https
(if HTTPS is used to connect to JOC Cockpit)JSON
MIME::Base64
- Cwd
Report Script per Agent
check_js7_agent.pl
calls an optional scriptreport_js7_agent.pl
per Agent.- This script does not connect to JOC Cockpit or perform any checks but simply serves to create individual notifications per Agent to the System Monitor.
- The script is executed if it is available from the same directory as the calling script and has to be executable.
- The script is parameterized to transfer the message type and notification to the System Monitor.
report_js7_agent.pl <joc-cockpit-url> <controller-id> <agent-url> <agent-status>
<joc-cockpit-url>
is the URL that has been specified as a parameter to the scriptcheck_js7_agent.pl
<scheduler-id>
is the JobScheduler Master ID that has been specified as a parameter to the scriptcheck_js7_agent.pl
<agent-url>
is the URL identifying the Agent<agent-status>
is one of "RUNNING" , "UNREACHABLE", "TERMINATING", "UNKNOWN_AGENT"Example report_js7_agent.pl#!/usr/bin/env perl use strict; use warnings; my ($jocUrl, $controllerId, $agentUrl, $agentStateText) = @ARGV; # do something with $jocUrl, $controllerId, $agentUrl, $agentStateText