Skip to content

Cluster CPU Load

Measures 1-, 5-, and 15-minute CPU load averages as supplied by /proc/loadavg.

cpuLoad -w w1,w5,w15 -c c1,c5,c15

The load average is a count of jobs in the run queue that are either runnable or waiting for disk I/O. The warning and critical limits are expressed as floating point triples for 1-, 5-, and 15-minute load averages. The limits only make sense with respect to core count on the machine. For example, a limit of 1.0 on an 8-core machine translates into a load average of 8, meaning the run queue length is the same as the core count and the machine is considered fully loaded.

CPU Load OK Nodes: 4

The detail section displays this information:

Detail Description
Status Status of the service
Status Information See following example for details
Last Check Time Date and time the CPU load status was checked
Next Scheduled Check Date and time of the next CPU load status check
Last State Change Date and time when the status last changed
In Schedule Downtime? Displays whether the warehouse is currently in scheduled downtime
Last Update Date and time when the service was last updated

The following detail is an Actian Warehouse instance with 1 leader and 3 workers each, configured with 32 cores. Warning alarm thresholds are 2.0, 1.25 and 1.25 for the 1-, 5-, and 15-minute load averages, respectively. Critical alarm threshold is 4.0 for 1, 5, and 15 load averages. A warning alarm would be triggered if the 1-minute load average exceeds 64 (32 cores * 2.0) or the 5- and 15-minute load averages exceed 40 (32 cores * 1.5).

cpu_load > 2.0,1.25,1.25 -> Warning, > 4.0,4.0,4.0 -> Critical Load Average

Role  Cores  1 min.   5 min.  15 min.   Node

   L    32    5.89     3.70     2.04    warehouses-av-8q99p2tn04of-vectorh-leader-0.warehouses-av-8q99p2tn04of-vectorh-leader.warehouses.svc.cluster.local

   W    32    4.47     2.55     1.36    warehouses-av-8q99p2tn04of-vectorh-worker-1.warehouses-av-8q99p2tn04of-vectorh-worker.warehouses.svc.cluster.local

   W    32    5.52     3.62     1.81    warehouses-av-8q99p2tn04of-vectorh-worker-0.warehouses-av-8q99p2tn04of-vectorh-worker.warehouses.svc.cluster.local

   W    32    4.53     3.00     1.50    warehouses-av-8q99p2tn04of-vectorh-worker-2.warehouses-av-8q99p2tn04of-vectorh-worker.warehouses.svc.cluster.local

Graph metrics:

leader_load5 leader node 5 minute load average
worker_load5_min min 5 minute load average for a slave node
worker_load5_avg average 5 minute load average for all slave nodes
worker_load5_max max 5 minute load average for a slave node