Details
-
Task
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
-
None
Description
sla_probe_hosts
Usage: sla_probe_hosts -cluster=cluster percentage duration {-hosts=hosts | --host_file=filename}
Lists all affected jobs in the following column format:
HOST JOB PREDICTION PREDICTED_SLA EXPECTED_IN_SLA_AFTER
where:
HOST - host being probed.
JOB - job that has tasks running on the host being probed.
PREDICTION - whether the job is going to stay within the provided SLA if the host is shut down.
PREDICTED_SLA - predicted effective percentage of up tasks if the host is shut down.
EXPECTED_IN_SLA_AFTER - expected wait time for the job to reach requested SLA threshold.
--cluster:
Aurora cluster name.
percentage:
Percentage of tasks required to be up within the duration.
duration:
Time interval (now - value) for the percentage of up tasks. Format: <value>
.
--hosts:
Comma-separated list of host names to probe.
--host_file:
Filename to read host names from (one per line).
Examples:
sla_probe_hosts --cluster=cluster 85 10mins --host_file=~/hosts.txt
Example :
aurora_admin sla_probe_hosts --cluster=cluster --hosts=host1 95 2hrs
host1 role/prod/job1 NO 92 12mins
host1 role/prod/job2 YES 96
host1 role/prod/job3 NO 93.5 45mins