On Thursday, 5 November 2020 14:36:31 UTC, Evelyn Pereira Souza wrote: > > Is this correct? 2 of 4 golden singals can't be done with blackbox?
I'd say not even that. Suppose you run blackbox exporter with a scrape interval of 5 seconds - you'll only get 12 measurements per minute. However your real application may be handling 1000's of requests per second, and these are *real* user requests performing real work. What you care about is the proportion of those real user requests which give errors, and the latency of those real user requests. These measurements are *much* more useful than synthetic measurements from blackbox exporter. Once you've got those real error and latency measurements, then the "traffic" measurement usually falls out in some application-specific way. e.g. for a web server it could be total number of HTTP requests. Resource saturation usually depends on some way of measuring the underlying resources, e.g. node_exporter for server resources (CPU, RAM, disk IOPS etc) blackbox_exporter is mainly useful for testing systems that you don't control, or which don't produce their own logs or metrics - a DNS server is a typical example, or a network link - or for alerting you quickly when a service has gone down completely. -- 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 on the web visit https://groups.google.com/d/msgid/prometheus-users/c986d5e0-13ac-4159-96f4-94a55dd015ado%40googlegroups.com.
