summaryrefslogtreecommitdiffstats
path: root/health/health.d/cpu.conf
diff options
context:
space:
mode:
Diffstat (limited to 'health/health.d/cpu.conf')
-rw-r--r--health/health.d/cpu.conf67
1 files changed, 67 insertions, 0 deletions
diff --git a/health/health.d/cpu.conf b/health/health.d/cpu.conf
new file mode 100644
index 0000000..ad69528
--- /dev/null
+++ b/health/health.d/cpu.conf
@@ -0,0 +1,67 @@
+
+# you can disable an alarm notification by setting the 'to' line to: silent
+
+ template: 10min_cpu_usage
+ on: system.cpu
+ class: Utilization
+ type: System
+component: CPU
+ os: linux
+ hosts: *
+ lookup: average -10m unaligned of user,system,softirq,irq,guest
+ units: %
+ every: 1m
+ warn: $this > (($status >= $WARNING) ? (75) : (85))
+ crit: $this > (($status == $CRITICAL) ? (85) : (95))
+ delay: down 15m multiplier 1.5 max 1h
+ info: average CPU utilization over the last 10 minutes (excluding iowait, nice and steal)
+ to: sysadmin
+
+ template: 10min_cpu_iowait
+ on: system.cpu
+ class: Utilization
+ type: System
+component: CPU
+ os: linux
+ hosts: *
+ lookup: average -10m unaligned of iowait
+ units: %
+ every: 1m
+ warn: $this > (($status >= $WARNING) ? (20) : (40))
+ crit: $this > (($status == $CRITICAL) ? (40) : (50))
+ delay: down 15m multiplier 1.5 max 1h
+ info: average CPU iowait time over the last 10 minutes
+ to: sysadmin
+
+ template: 20min_steal_cpu
+ on: system.cpu
+ class: Latency
+ type: System
+component: CPU
+ os: linux
+ hosts: *
+ lookup: average -20m unaligned of steal
+ units: %
+ every: 5m
+ warn: $this > (($status >= $WARNING) ? (5) : (10))
+ crit: $this > (($status == $CRITICAL) ? (20) : (30))
+ delay: down 1h multiplier 1.5 max 2h
+ info: average CPU steal time over the last 20 minutes
+ to: sysadmin
+
+## FreeBSD
+ template: 10min_cpu_usage
+ on: system.cpu
+ class: Utilization
+ type: System
+component: CPU
+ os: freebsd
+ hosts: *
+ lookup: average -10m unaligned of user,system,interrupt
+ units: %
+ every: 1m
+ warn: $this > (($status >= $WARNING) ? (75) : (85))
+ crit: $this > (($status == $CRITICAL) ? (85) : (95))
+ delay: down 15m multiplier 1.5 max 1h
+ info: average CPU utilization over the last 10 minutes (excluding nice)
+ to: sysadmin