I'm working on a Zabbix implementation. The customer has an extended service catalog (description of all IT services to deliver to the end-user) and also a high number of nodes to monitor (over 800).
I would like to offer our business managers a status view on our delivered IT services. So I looked into the 'IT services' part of Zabbix, but I can't get my way around this yet.
What I would like is the following tree:
etc. etc.
However, in this case, service 1 (for example) is relying on more than one node for each location. For example: if one of our compute cluster nodes is down, we still have 20 others left and the IT service is still being delivered.
I'm having trouble to connect the trigger to the service, if my service covers more than one machine. Of course I can create a custom trigger, but I wonder how to manage the 'coming and going' of hosts in this particular service group. Do I have to change the custom trigger manually in case I pull out one machine of the cluster?
Righ now, I think my best option is to write a small script/extension that monitors the status of cluster members, through the cluster master, but this creates a SPOF and is not always possible.
Is there anyone with some experiences in this area, willing to provide some advise?
I would like to offer our business managers a status view on our delivered IT services. So I looked into the 'IT services' part of Zabbix, but I can't get my way around this yet.
What I would like is the following tree:
- location 1
- service 1
- service 2
- service 3
- service 4
- service 5
- location 2
- service 1
- service 2
- service 3
- service 4
- service 5
- location 3
- service 1
- service 2
- service 3
etc. etc.
However, in this case, service 1 (for example) is relying on more than one node for each location. For example: if one of our compute cluster nodes is down, we still have 20 others left and the IT service is still being delivered.
I'm having trouble to connect the trigger to the service, if my service covers more than one machine. Of course I can create a custom trigger, but I wonder how to manage the 'coming and going' of hosts in this particular service group. Do I have to change the custom trigger manually in case I pull out one machine of the cluster?
Righ now, I think my best option is to write a small script/extension that monitors the status of cluster members, through the cluster master, but this creates a SPOF and is not always possible.
Is there anyone with some experiences in this area, willing to provide some advise?
Comment