Snmp-interface perfdata not saving for CentOS 8

I’ve got snmp-interface monitors on a bunch of Linux boxes, and I’m using php4nagios to graph the results. For some new CentOS 8 boxes, I’m not getting graphs saved, though the plugin is returning information. RHEL6 and 7 have worked just fine for years. Has anyone else had this problem?

Output from a 7.x server looks like this:
eth0:UP (0.0Mbps/0.0Mbps), eth1:UP (0.0Mbps/0.0Mbps):2 UP: OK | 'eth0_in_bps'=1505;60000000;80000000;0;1000000000 'eth0_out_bps'=17358;40000000;50000000;0;1000000000 'eth1_in_bps'=0;60000000;80000000;0;1000000000 'eth1_out_bps'=0;40000000;50000000;0;1000000000

Compared to 8.0, the only difference I see is a 0 in the last value:
eth0:UP (0.0Mbps/0.1Mbps), eth1:UP (0.0Mbps/0.0Mbps):2 UP: OK | 'eth0_in_bps'=16910;60000000;80000000;0;0 'eth0_out_bps'=72539;40000000;50000000;0;0 'eth1_in_bps'=25402;60000000;80000000;0;0 'eth1_out_bps'=6293;40000000;50000000;0;0

The command being run in both cases is:
'/usr/lib64/nagios/plugins/contrib/manubulon-snmp/plugins/' '-d' '300' '-x' '*****' '-w' '60,40' '-t' '5' '-n' 'eth' '-l' '*****' '-k' '-f' '--64bits' '-c' '80,50' '-Y' '-X' '*****' '-M' '-L' 'sha,aes' '-H' 'host1' '-B'

Thanks for any suggestions!

It looks like the max is 1000000000 with CentOS 7 and is 0 with CentOS 8. Maybe the version of the plugin changed the default behavior? If the max is 0 for the graph, I guess it doesn’t know how to draw it with a value outside of that range? I haven’t used PNP4Nagios in a long time, migrating to carbon/graphite/grafana.

Yes, that seems to be the case. Something must have changed in net-snmp between 5.7 and 5.8, or something different about how RHEL 8 handles interfaces.

$ snmpwalk host1
IF-MIB::ifSpeed.1 10000000
IF-MIB::ifSpeed.2 1000000000
IF-MIB::ifSpeed.3 1000000000
$ snmpwalk host2
IF-MIB::ifSpeed.1 10000000
IF-MIB::ifSpeed.2 0
IF-MIB::ifSpeed.3 0