Varnish exporter for Prometheus
main_n_ban
grouping. Metrics will now have individual bans_<type>
metrics instad of the grouped metric that had type
as a label. (#51 @glennslaven)
a
metrics as gauges instead of counters. (#48 @glennslaven)-test
mode to wait for full metrics scrape before continuing.-no-exit
. Default behavior is now not to exit on scrape errors as it should be for a long running HTTP server.
-no-exit
but the process behaves as before.-exit-on-errors
has been added for users who want the old default behavior back.varnishstat
output type
property.github.com/prometheus/client_golang
v1.0.0-docker-container-name
to signal that varnishstat
should be ran in a docker container with docker exec <container-name>
.-with-go-metrics
. Before they were included by default. Now dropped to export less clutter that majority of users will never need (@nipuntalukdar).varnish_backend_up
with Varnish 4.0 and earlier versions.varnish_backend_up
with 1/0 value that reflects the latest health probe result. The Varnish bitmap uint64 varnish_backend_happy
as a prometheus float metric was not that useful in detecting latest up/down per backend.-varnishstat-path
(by @zstyblik)-no-exit
mode that does not exit the process if varnish is not running at the time of startup.type
and ident
properties from varnishstat JSON output. If ident
is not present, it is now parsed from the metric name.