I want to pull values from suricata stats into Zabbix. An example of what these look like are below.
I'm trying to understand the value, benefit of the items actual name. For example one naming convention could be
suricata.stats.STAT.NAME
specifically
suricata.stats.capture.kernel_drops
Another strategy could be
suricata.stats[STAT.NAME]
or
suricata.stats[capture.kernel_drops]
Is there a reason why I might chose one form vs the other? Is there another form to consider?
Does one of these strategies have an impact on storage? Retrieval? Expiration? Graphing?
capture.kernel_packets | Total | 3391904
capture.kernel_drops | Total | 486
capture.kernel_ifdrops | Total | 0
decoder.pkts | Total | 3391418
decoder.bytes | Total | 2414269170
decoder.invalid | Total | 1
decoder.ipv4 | Total | 3379115
decoder.ipv6 | Total | 2636
decoder.ethernet | Total | 3391418
decoder.arp | Total | 9398
decoder.unknown_ethertype | Total | 269
decoder.chdlc | Total | 0
decoder.raw | Total | 0
decoder.null | Total | 0
decoder.sll | Total | 0
decoder.tcp | Total | 1400986
tcp.syn | Total | 31126
tcp.synack | Total | 28924
tcp.rst | Total | 3037
decoder.udp | Total | 1978834
capture.kernel_drops | Total | 486
capture.kernel_ifdrops | Total | 0
decoder.pkts | Total | 3391418
decoder.bytes | Total | 2414269170
decoder.invalid | Total | 1
decoder.ipv4 | Total | 3379115
decoder.ipv6 | Total | 2636
decoder.ethernet | Total | 3391418
decoder.arp | Total | 9398
decoder.unknown_ethertype | Total | 269
decoder.chdlc | Total | 0
decoder.raw | Total | 0
decoder.null | Total | 0
decoder.sll | Total | 0
decoder.tcp | Total | 1400986
tcp.syn | Total | 31126
tcp.synack | Total | 28924
tcp.rst | Total | 3037
decoder.udp | Total | 1978834
suricata.stats.STAT.NAME
specifically
suricata.stats.capture.kernel_drops
Another strategy could be
suricata.stats[STAT.NAME]
or
suricata.stats[capture.kernel_drops]
Is there a reason why I might chose one form vs the other? Is there another form to consider?
Does one of these strategies have an impact on storage? Retrieval? Expiration? Graphing?
Name should be more human readable than key for example... a (dependent) item key like "suricata.stats.capture.kernel_drops" looks just fine but Name .. Maybe simple "Kernel drops" is more readable... Name also appears in graph legends, so it may be easier to read graphs later. so yea.. for graphing it may have influence... for storage or retrieval or expiration... no
Comment