We use Munin to gather statistics from our Tor exits. For the configuration of munin-node on the exits (“clientside munin-node”), see munin_resource_monitoring.

This is the “serverside munin” configuration. The only interesting part might be the aggregate bandwidth graph. We don't update this page every time we add or remove a relay, so don't expect everything to match the current configuration.

TODO We could use some more fancy rules to warn when throughput is way below average, ie. when something is wrong with the relay.


includedir /etc/munin/munin-conf.d

contacts me mail -s "Munin notification ${var:host}" NICEEMAIL warning critical

    use_node_name yes

    use_node_name yes
# warning when below 50 mbit/s, critical when below 10 mbit/s
    if_venet0.up.warning  10000000:50000000
    if_venet0.up.critical        0:10000000
# ... similar entries for all the other relays ...

    update no
    total_bandwidth.graph_args --base 1000 -l 0
    total_bandwidth.graph_category Network
    total_bandwidth.graph_title Aggregated bandwidth
    total_bandwidth.upload.label upload
    total_bandwidth.update no yes
    total_bandwidth.upload.stack \ \ \ \ \ \ \ \ \ \ \ \
    total_bandwidth.upload.type COUNTER