[Bug]: Error in packet, Reason: (tooBig) Response message would have been too large on MikroTik Switches

[Beleggrodion]

Please try to fill out as much of the information below as you can. Thank you!

• Yes, I've searched similar issues on GitHub and didn't find any.

Which version contains the bug?

1.4.4

Describe the bug

When i try to check a mikrotik switch, ex with: check_interfaces -h 192.168.x.x -c public i receive the following error:

Error in packet
Reason: (tooBig) Response message would have been too large.

On our old cisco switch all work as expected. I tested it with different models.

How to recreate the bug?

Enable snmp on the switch and do a request from monitoring host.

[RincewindsHat]

Hi @Beleggrodion,
thanks for reporting this. It is kinda hard to debug without having a mikrotik switch.
Could you try adding --mode nonbulk to the command?
My current guess is, that the switch does not want to send too much data in a single message (but I might be wrong).
Logs from the switch might be helpful, if available.

[Beleggrodion]

Ok, with --mode nonbulk it looks like to work

Below the log when i receive the error above from one of the switch logs:

 2025-06-13 11:19:50 snmp *: packet(v2) from: 192.168.x.x community: public
 2025-06-13 11:19:50 snmp,debug *: error: max repetitions exceeded

So i tested it with the --max-repetitions Parameter. The max which works is 50, with 51 i receive the mentioned message.

But it looks like, i also cannot use this check script for one of our core switch at a customer side :( Because we use Mikrotik capsman the central manage all the ap's and all ap's return also their wifi interface statistics to the core switch, so we had for each ssid two interfaces (2.4 ghz and 5 ghz) and we had multiple ssid (intern, guests, intern-limited, etc.). I think i need here bash scripts which call all oids directly :/ I don't find any settings to limit the return on the switch. Because single request of one interface tooks more than 43 secs.

[RincewindsHat]

Sounds to me like the table of interfaces is far too big then. I guess a specialised Monitoring Plugin might be more suitable then.
Otherwise one would have to implement some kind of limiter in this one.