Does anyone have the same problem that a device ( from Honeywell) has somtimes not a a special OID an 30 min later has the OID and loosing a diffrent one?
this looks very unusual. You should talk to the vendor of the device about the SNMP agent implementation.
Checkmk is just doing snmpbulkwalks. If that is not supported, you may change the settings with the rulesets “Bulk walk: Hosts using bulk walk”, “Hosts not using Inline-SNMP” and/or “Timing settings for SNMP access”.
A lot of “appliance” devices aren’t meant for their local CPU to do a lot of “outside work” (like handling your snmp request). Sometimes, you can make things more stable by increasing the interval between checks. Sometimes, also, you may know of stressful loads being placed on an appliance a certain time periods, so a time period definition and what you do based on that can also help.
Of course, sometimes, you’ve already got “skunk works” or other monitors attempting snmp against the appliance already… and so it was marginally “keeping up” with the requests and you’re seeing times when a whole lot of requests are coming to it at once periodically and the end appliance can’t manage it all.