Skip to content

Latest commit

 

History

History
173 lines (147 loc) · 5.16 KB

README.md

File metadata and controls

173 lines (147 loc) · 5.16 KB

check_orchestrator

Go

This repository contains a Nagios / Icinga check to monitor Orchestrator.

This check was written by Matthias Crauwels [email protected] at Ghent University. It is published with an MIT license

Usage

$ bin/check_orchestrator
Usage:
check_orchestrator [subcommand] [OPTIONS]
SubCommands:
clusterhealth
clusterinfo
status

Commands

status

Usage

$ bin/check_orchestrator status -h
Usage:
check_orchestrator status [OPTIONS]
Application Options:
-H, --host= Hostname (default: localhost)
-p, --port= Port (default: 3000)
-S, --ssl Use SSL
-I, --insecure Do not check SSL cert
-U, --uri= URI (default: api/health)
--http-auth-name Name for http auth 
--http-auth-password Password for http auth  
Help Options:
-h, --help Show this help message

Success

$ bin/check_orchestrator status
ORCHESTRATOR_STATUS OK: Application node is healthy

Errors

$ bin/check_orchestrator status
ORCHESTRATOR_STATUS CRITICAL: Application node is unhealthy dial tcp 127.0.0.1:20192: getsockopt: connection refused

clusterinfo

Usage

$ bin/check_orchestrator clusterinfo -h
Usage:
check_orchestrator clusterinfo [OPTIONS]
Application Options:
-H, --host= Hostname (default: localhost)
-p, --port= Port (default: 3000)
-S, --ssl Use SSL
-I, --insecure Do not check SSL cert
-U, --uri= URI (default: api/clusters-info)
--http-auth-name Name for http auth 
--http-auth-password Password for http auth 
Help Options:
-h, --help Show this help message

Success

$ bin/check_orchestrator clusterinfo
ORCHESTRATOR_CLUSTERINFO OK: This instance manages following clusters: 127.0.0.1:20192 (HasAutomatedMasterRecovery = false) (HasAutomtedIntermediateMasterRecovery = false), localhost:20192 (HasAutomatedMasterRecovery = false) (HasAutomtedIntermediateMasterRecovery = false)

Errors

  • Orchestrator has no clusters configured
 $ /tmp/check_orchestrator clusterinfo
ORCHESTRATOR_CLUSTERINFO WARNING: This Orchestrator is responding correctly but is not managing any clusters.

clusterhealth

Usage

$ bin/check_orchestrator clusterhealth --help
Usage:
check_orchestrator clusterhealth --alias=<clusteralias> [OPTIONS]
Application Options:
-a, --alias= ClusterAlias
-H, --host= Hostname (default: localhost)
-p, --port= Port (default: 3000)
-S, --ssl Use SSL
-I, --insecure Do not check SSL cert
-t, --timeout= Timeout for SecondsSinceLastSeen (default: 300)
-w, --lag-warning= Slave lag warning threshold (default: 300)
-c, --lag-critical= Slave lag critical threshold (default: 600)
--http-auth-name Name for http auth 
--http-auth-password Password for http auth 
Help Options:
-h, --help Show this help message

Success

$ bin/check_orchestrator clusterhealth --alias=127.0.0.1:20192
ORCHESTRATOR_CLUSTERHEALTH OK: Cluster 127.0.0.1:20192 is doing OK

Errors

  • no alias
$ bin/check_orchestrator clusterhealth
the required flag `-a, --alias' was not specified
  • multiple writers
$ bin/check_orchestrator clusterhealth --alias=127.0.0.1:20192
ORCHESTRATOR_CLUSTERHEALTH CRITICAL: [SPLIT BRAIN] There are 2 writable servers in cluster 127.0.0.1:20192
  • Slave thread(s) not running
$ bin/check_orchestrator clusterhealth --alias=127.0.0.1:20192
ORCHESTRATOR_CLUSTERHEALTH CRITICAL: In cluster 127.0.0.1:20192 the Slave_IO-thread is not running on host 127.0.0.1:20195
$ bin/check_orchestrator clusterhealth --alias=127.0.0.1:20192
ORCHESTRATOR_CLUSTERHEALTH CRITICAL: In cluster 127.0.0.1:20192 the Slave_SQL-thread is not running on host 127.0.0.1:20195
  • Slave lag
$ bin/check_orchestrator clusterhealth --alias 127.0.0.1:20192 -w 30 -c 60
ORCHESTRATOR_CLUSTERHEALTH WARNING: In cluster 127.0.0.1:20192 host 127.0.0.1:20195 is 53 seconds lagging (warning threshold 30)

 $ bin/check_orchestrator clusterhealth --alias 127.0.0.1:20192 -w 30 -c 60
ORCHESTRATOR_CLUSTERHEALTH CRITICAL: In cluster 127.0.0.1:20192 host 127.0.0.1:20195 is 65 seconds lagging (critical threshold 60)

note the clusterhealth-command take the downtime setting in Orchestrator into account...

Nagios Installation

Configuration

Assuming a standard installation of Nagios, the plugin can be executed from the machine that Nagios is running on.

cp check_orchestrator /usr/local/nagios/libexec/plugins/check_orchestrator
chmod +x /usr/local/nagios/libexec/plugins/check_orchestrator

Add the following service definition to your server config:

define service {
        use                             local-service
        host_name                       localhost
        service_description             <command_description>
        check_command                   <command_name>
        }

Add the following command definition to your commands config (commands.config):

define command{
        command_name    <command_name>
        command_line    /usr/local/nagios/libexec/plugins/check_orchestrator <command> <parameters>
        }

More info about options in Commands.