Current VIRL bootstrap.sh simply load-balances across all three VIRL hosts.
This JIRA is to implement usage and status monitoring, as well as a more granular selection of VIRL hosts:
- Monitor the health state of each VIRL host (as per criteria tbd). Take a VIRL host out of production, and notify the administrator, if a host is deemed "not healthy".
- Monitor usage. Be able to identify number of tests run (successful/failed) per host, per day, with historic data. Be able to quickly see number of jobs currently run, and high watermarks in terms of CPU, memory and disk utilization
- Direct jobs to specific VIRL hosts. Be able to take a VIRL host out of production, and set it into a "testing"/"staging" mode where infrastructure changes can be tested and jobs can be deliberately directed to this host while other (production) jobs will be load-shared across the other hosts