summaryrefslogtreecommitdiffstats
path: root/health/health.d/net.conf
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--health/health.d/net.conf165
1 files changed, 165 insertions, 0 deletions
diff --git a/health/health.d/net.conf b/health/health.d/net.conf
new file mode 100644
index 0000000..ae3c26e
--- /dev/null
+++ b/health/health.d/net.conf
@@ -0,0 +1,165 @@
+
+# you can disable an alarm notification by setting the 'to' line to: silent
+
+# -----------------------------------------------------------------------------
+# net traffic overflow
+
+ template: interface_speed
+ on: net.net
+ os: *
+ hosts: *
+ families: *
+ calc: ( $nic_speed_max > 0 ) ? ( $nic_speed_max) : ( nan )
+ units: Mbit
+ every: 10s
+ info: The current speed of the physical network interface
+
+ template: 1m_received_traffic_overflow
+ on: net.net
+ os: linux
+ hosts: *
+ families: *
+ lookup: average -1m unaligned absolute of received
+ calc: ($interface_speed > 0) ? ($this * 100 / ($interface_speed * 1000)) : ( nan )
+ units: %
+ every: 10s
+ warn: $this > (($status >= $WARNING) ? (80) : (85))
+ crit: $this > (($status == $CRITICAL) ? (85) : (90))
+ delay: down 1m multiplier 1.5 max 1h
+ info: interface received bandwidth usage over net device speed max
+ to: sysadmin
+
+ template: 1m_sent_traffic_overflow
+ on: net.net
+ os: linux
+ hosts: *
+ families: *
+ lookup: average -1m unaligned absolute of sent
+ calc: ($interface_speed > 0) ? ($this * 100 / ($interface_speed * 1000)) : ( nan )
+ units: %
+ every: 10s
+ warn: $this > (($status >= $WARNING) ? (80) : (85))
+ crit: $this > (($status == $CRITICAL) ? (85) : (90))
+ delay: down 1m multiplier 1.5 max 1h
+ info: interface sent bandwidth usage over net device speed max
+ to: sysadmin
+
+# -----------------------------------------------------------------------------
+# dropped packets
+
+# check if an interface is dropping packets
+# the alarm is checked every 1 minute
+# and examines the last 10 minutes of data
+
+template: inbound_packets_dropped
+ on: net.drops
+ os: linux
+ hosts: *
+families: *
+ lookup: sum -10m unaligned absolute of inbound
+ units: packets
+ every: 1m
+ warn: $this >= 5
+ delay: down 1h multiplier 1.5 max 2h
+ info: interface inbound dropped packets in the last 10 minutes
+ to: sysadmin
+
+template: outbound_packets_dropped
+ on: net.drops
+ os: linux
+ hosts: *
+families: *
+ lookup: sum -10m unaligned absolute of outbound
+ units: packets
+ every: 1m
+ warn: $this >= 5
+ delay: down 1h multiplier 1.5 max 2h
+ info: interface outbound dropped packets in the last 10 minutes
+ to: sysadmin
+
+template: inbound_packets_dropped_ratio
+ on: net.packets
+ os: linux
+ hosts: *
+families: *
+ lookup: sum -10m unaligned absolute of received
+ calc: (($inbound_packets_dropped != nan AND $this > 0) ? ($inbound_packets_dropped * 100 / $this) : (0))
+ units: %
+ every: 1m
+ warn: $this >= 0.1
+ crit: $this >= 2
+ delay: down 1h multiplier 1.5 max 2h
+ info: the ratio of inbound dropped packets vs the total number of received packets of the network interface, during the last 10 minutes
+ to: sysadmin
+
+template: outbound_packets_dropped_ratio
+ on: net.packets
+ os: linux
+ hosts: *
+families: *
+ lookup: sum -10m unaligned absolute of sent
+ calc: (($outbound_packets_dropped != nan AND $this > 0) ? ($outbound_packets_dropped * 100 / $this) : (0))
+ units: %
+ every: 1m
+ warn: $this >= 0.1
+ crit: $this >= 2
+ delay: down 1h multiplier 1.5 max 2h
+ info: the ratio of outbound dropped packets vs the total number of sent packets of the network interface, during the last 10 minutes
+ to: sysadmin
+
+
+# -----------------------------------------------------------------------------
+# FIFO errors
+
+# check if an interface is having FIFO
+# buffer errors
+# the alarm is checked every 1 minute
+# and examines the last 10 minutes of data
+
+template: 10min_fifo_errors
+ on: net.fifo
+ os: linux
+ hosts: *
+families: *
+ lookup: sum -10m unaligned absolute
+ units: errors
+ every: 1m
+ warn: $this > 0
+ delay: down 1h multiplier 1.5 max 2h
+ info: interface fifo errors in the last 10 minutes
+ to: sysadmin
+
+
+# -----------------------------------------------------------------------------
+# check for packet storms
+
+# 1. calculate the rate packets are received in 1m: 1m_received_packets_rate
+# 2. do the same for the last 10s
+# 3. raise an alarm if the later is 10x or 20x the first
+# we assume the minimum packet storm should at least have
+# 10000 packets/s, average of the last 10 seconds
+
+template: 1m_received_packets_rate
+ on: net.packets
+ os: linux freebsd
+ hosts: *
+families: *
+ lookup: average -1m unaligned of received
+ units: packets
+ every: 10s
+ info: the average number of packets received during the last minute
+
+template: 10s_received_packets_storm
+ on: net.packets
+ os: linux freebsd
+ hosts: *
+families: *
+ lookup: average -10s unaligned of received
+ calc: $this * 100 / (($1m_received_packets_rate < 1000)?(1000):($1m_received_packets_rate))
+ every: 10s
+ units: %
+ warn: $this > (($status >= $WARNING)?(200):(5000))
+ crit: $this > (($status >= $WARNING)?(5000):(6000))
+options: no-clear-notification
+ info: the % of the rate of received packets in the last 10 seconds, compared to the rate of the last minute (clear notification for this alarm will not be sent)
+ to: sysadmin