Hello everyone,
I know that in the past it was not possible to generate Piggyback data from an SNMP-based check. Therefore my question is: has this changed in the meantime?
The background is a WLAN controller cluster. Via SNMP, I can see both the access points managed locally by a controller as well as access points managed by other cluster members.
For locally managed APs, a lot of interesting information is available, such as:
- Status
- Number of connected clients
- Radio parameters
- Channel and transmit power information
- SSIDs
- etc.
For APs managed by another cluster member, I can still retrieve basic information such as name, MAC address, IP address, and the controller currently managing the AP.
My initial approach was to monitor the APs as services on the respective WLC. However, I would actually prefer to model the APs as individual hosts in Checkmk – similar to the VMware/vCenter integration, where VMs appear as separate hosts and receive their data via Piggyback. Ideally, I would also like to create the AP hosts automatically via DCD.
A Special Agent would of course be an option. However, that would mean implementing all SNMP communication myself and doing the “heavy lifting”, even though Checkmk’s SNMP engine is already collecting the required data very efficiently. I would prefer to avoid that. In addition, whenever an AP changes its controller, the cluster would likely need to be rediscovered, which is not particularly appealing either.
What I would really like is to treat the APs as independent objects, regardless of which controller is currently managing them. In that model, the controller would merely act as the data source rather than being the actual monitored host.
Has anyone implemented something similar before or can point me in the right direction?
Any ideas or recommendations would be greatly appreciated.
Daniel