What is the best practice for a jenkins hearbeat script? -
i have multiple jenkins instances running not have slave/master relationship. best way have them monitor each other , warn admin if 1 instance goes down. current approach using have job on instance triggers job on instance b, in turn checks resources of instance b healthy. instance b performs same trigger process on instance a. setup? best way trigger job , output status?
is setup?
if works use case, setup; is, don't find tweaking scripts frequently, or being falsely alarmed.
there may cases using job queue problematic - e.g. there many jobs execute before one, or long-running ones block heartbeat processed. then, perhaps possible keep 1 job executor free perform jobs heartbeat; 1 (job restrictions) seems fit.
additionally, coupling heartbeats sequentially means after first failure, there no more information other nodes in sequence; when instance down, don't know state of instance b.
i'd check server monitoring software: nagios, cacti etc. you'll find scripts integrate jenkins it, , looks integrating monitoring jobs easy; here on so, memangesh wrote script checks whether particular jenkins job url can accessed via http.
using such software has additional benefit hints @ on servers - e.g. when hard disk partitions run full, when server's load running high, etc.
Comments
Post a Comment