organize

docker-compose.yml

@@ -12,7 +12,7 @@ services:
     build:
       dockerfile: Dockerfile
     container_name: nagios
-    image: docker-registry1.in.thelinuxpro.net:5000/tlp/nagios:241228.0.3
+    image: docker-registry1.in.thelinuxpro.net:5000/tlp/nagios:241230.0.1
     networks:
       infra_dev_net:
         ipv4_address: 10.99.23.36

overlay/opt/nagios/etc/objects/001-templates.cfg

@@ -67,7 +67,7 @@ define host{
 	name				linux-server	; The name of this host template
 	use				generic-host	; This template inherits other values from the generic-host template
 	check_period			24x7		; By default, Linux hosts are checked round the clock
-	check_interval			5		; Actively check the host every 5 minutes
+	check_interval			2		; Actively check the host every 5 minutes
 	retry_interval			1		; Schedule host check retries at 1 minute intervals
 	max_check_attempts		10		; Check each Linux host 10 times (max)
         check_command           	check-host-alive ; Default command to check Linux hosts
@@ -80,43 +80,22 @@ define host{
 	register			0		; DONT REGISTER THIS DEFINITION - ITS NOT A REAL HOST, JUST A TEMPLATE!
 	}
 
-
+### Host Template
-
-# Windows host definition template - This is NOT a real host, just a template!
-
 define host{
-	name			windows-server	; The name of this host template
+	name			docker-host
-	use			generic-host	; Inherit default values from the generic-host template
+	use			generic-host
-	check_period		24x7		; By default, Windows servers are monitored round the clock
+	hostgroups		docker-hosts
-	check_interval		5		; Actively check the server every 5 minutes
+	check_period		24x7
-	retry_interval		1		; Schedule host check retries at 1 minute intervals
+	check_interval		3
-	max_check_attempts	10		; Check each server 10 times (max)
+	retry_interval		1
-	check_command		check-host-alive	; Default command to check if servers are "alive"
+	max_check_attempts	3
-	notification_period	24x7		; Send notification out at any time - day or night
+	check_command		check-host-alive
-	notification_interval	30		; Resend notifications every 30 minutes
+	notification_period	24x7
-	notification_options	d,r		; Only send notifications for specific host states
+	notification_interval	120
-	contact_groups		admins		; Notifications get sent to the admins by default
+	notification_options	d,u,r
-	hostgroups		windows-servers ; Host groups that Windows servers should be a member of
+	contact_groups		admins
-	register		0		; DONT REGISTER THIS - ITS JUST A TEMPLATE
+	register		0
-	}
+}
-
-
-# We define a generic printer template that can be used for most printers we monitor
-
-define host{
-	name			generic-printer	; The name of this host template
-	use			generic-host	; Inherit default values from the generic-host template
-	check_period		24x7		; By default, printers are monitored round the clock
-	check_interval		5		; Actively check the printer every 5 minutes
-	retry_interval		1		; Schedule host check retries at 1 minute intervals
-	max_check_attempts	10		; Check each printer 10 times (max)
-	check_command		check-host-alive	; Default command to check if printers are "alive"
-	notification_period	workhours		; Printers are only used during the workday
-	notification_interval	30		; Resend notifications every 30 minutes
-	notification_options	d,r		; Only send notifications for specific host states
-	contact_groups		admins		; Notifications get sent to the admins by default
-	register		0		; DONT REGISTER THIS - ITS JUST A TEMPLATE
-	}
 
 
 # Define a template for switches that we can reuse
@@ -164,8 +143,8 @@ define service{
         is_volatile                     0       		; The service is not volatile
         check_period                    24x7			; The service can be checked at any time of the day
         max_check_attempts              3			; Re-check the service up to 3 times in order to determine its final (hard) state
-        check_interval                  10			; Check the service every 10 minutes under normal conditions
+        check_interval                  2			; Check the service every 10 minutes under normal conditions
-        retry_interval                  2			; Re-check the service every two minutes until a hard state can be determined
+        retry_interval                  1			; Re-check the service every two minutes until a hard state can be determined
         contact_groups                  admins			; Notifications get sent out to everyone in the 'admins' group
 	notification_options		w,u,c,r			; Send notifications about warning, unknown, critical, and recovery events
         notification_interval           60			; Re-notify about service problems every hour

overlay/opt/nagios/etc/objects/101-docker-hosts.cfg

@@ -0,0 +1,28 @@
+### Hosts
+define host{
+	use			docker-host
+	host_name		pi501.in.thelinuxpro.net
+	alias			pi501
+	address			10.99.22.51
+}
+
+define host{
+	use			docker-host
+	host_name		pi502.in.thelinuxpro.net
+	alias			pi502
+	address			10.99.22.52
+}
+
+define host{
+	use			docker-host
+	host_name		pi503.in.thelinuxpro.net
+	alias			pi503
+	address			10.99.22.53
+}
+
+define host{
+	use			docker-host
+	host_name		pi504.in.thelinuxpro.net
+	alias			pi504
+	address			10.99.22.54
+}

overlay/opt/nagios/etc/objects/301-docker-services-hostgroup.cfg

@@ -0,0 +1,43 @@
+
+### Service definitions for all docker hosts
+define service{
+	use			generic-service,graphed-service
+	hostgroup_name		docker-hosts
+	service_description	PING
+	check_command		check_ping!100.0,20%!500.0,60%
+}
+
+define service{
+	use			generic-service,graphed-service
+	hostgroup_name		docker-hosts
+	service_description	Docker Daemon
+	check_command		check_nrpe!check_proc_docker
+}
+
+define service{
+	use			generic-service,graphed-service
+	hostgroup_name		docker-hosts
+	service_description	Disk: /
+	check_command		check_nrpe!check_disk
+}
+
+define service{
+	use			generic-service,graphed-service
+	hostgroup_name		docker-hosts
+	service_description	Disk: /var/lib/docker/volumes
+	check_command		check_nrpe!check_disk_docker
+}
+
+define service{
+	use			generic-service,graphed-service
+	hostgroup_name		docker-hosts
+	service_description	System: Load
+	check_command		check_nrpe!check_load
+}
+
+define service{
+	use			generic-service,graphed-service
+	hostgroup_name		docker-hosts
+	service_description	System: Total Processes
+	check_command		check_nrpe!check_procs
+}

overlay/opt/nagios/etc/objects/401-docker-services.cfg

@@ -1,114 +1,3 @@
-
-### Host Template
-define host{
-	name			docker-host
-	use			generic-host
-	hostgroups		docker-hosts
-	check_period		24x7
-	check_interval		5
-	retry_interval		1
-	max_check_attempts	3
-	check_command		check-host-alive
-	notification_period	24x7
-	notification_interval	120
-	notification_options	d,u,r
-	contact_groups		admins
-	register		0
-}
-
-### Hosts
-define host{
-	use			docker-host
-	host_name		pi501.in.thelinuxpro.net
-	alias			pi501
-	address			10.99.22.51
-}
-
-define host{
-	use			docker-host
-	host_name		pi502.in.thelinuxpro.net
-	alias			pi502
-	address			10.99.22.52
-}
-
-define host{
-	use			docker-host
-	host_name		pi503.in.thelinuxpro.net
-	alias			pi503
-	address			10.99.22.53
-}
-
-define host{
-	use			docker-host
-	host_name		pi504.in.thelinuxpro.net
-	alias			pi504
-	address			10.99.22.54
-}
-
-### Service definitions for all docker hosts
-define service{
-	use			local-service,graphed-service
-	hostgroup_name		docker-hosts
-	service_description	PING
-	check_command		check_ping!100.0,20%!500.0,60%
-}
-
-define service{
-	use			local-service,graphed-service
-	hostgroup_name		docker-hosts
-	service_description	Docker Daemon
-	check_command		check_nrpe!check_proc_docker
-}
-
-define service{
-	use			local-service,graphed-service
-	hostgroup_name		docker-hosts
-	service_description	Disk: /
-	check_command		check_nrpe!check_disk
-}
-
-define service{
-	use			local-service,graphed-service
-	hostgroup_name		docker-hosts
-	service_description	Disk: /var/lib/docker/volumes
-	check_command		check_nrpe!check_disk_docker
-}
-
-define service{
-	use			local-service,graphed-service
-	hostgroup_name		docker-hosts
-	service_description	System: Load
-	check_command		check_nrpe!check_load
-}
-
-define service{
-	use			local-service,graphed-service
-	hostgroup_name		docker-hosts
-	service_description	System: Total Processes
-	check_command		check_nrpe!check_procs
-}
-
-define service{
-	use			local-service,graphed-service
-	hostgroup_name		docker-hosts
-	service_description	System: Zombie Processes
-	check_command		check_nrpe!check_procs
-}
-
-define service{
-	use			local-service,graphed-service
-	hostgroup_name		docker-hosts
-	service_description	System: Logged In users
-	check_command		check_nrpe!check_users
-}
-
-define service{
-	use			local-service,graphed-service
-	hostgroup_name		docker-hosts
-	service_description	DNS Query: A Record
-	check_command		check_dns_a
-}
-
 ### Service Definitions per docker host
 define service{
 	use			local-service,graphed-service