Hello! I have snmp_exporter set up to scrape a number of network switches. I have a weird issue where a specific device is a lot slower than all the others of the same model to scrape - but only when using snmp_exporter!
The time needed for snmp_exporter to scrape the device(snmp_scrape_duration_seconds) ranges from 40 to 90+ seconds(the current timeout). However, using snmpbulkwalk from the same server, results in a much faster scrape, in my tests 4-10 seconds. I tried with and without the -Cr flag set to 10 which is the same value as I am using as max_repetitions for snmp_exporter, without change. I have also tried scraping the resolved IP directly to exclude DNS. My current theory is that there might be some behavior where snmp_exporter/gosnmp differs from what snmpbulkwalk does, and that this tickles the network switch in the wrong way. Anyone experienced something similar? Anyone have any ideas about other snmpbulkwalk flags I could try to replicate the issue? Best regards, Wilhelm -- You received this message because you are subscribed to the Google Groups "Prometheus Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/prometheus-users/62b0651b-9ae0-432b-8ce6-f4d5eaa2b85bn%40googlegroups.com.
