This PR changes they way of communication with remote jobs, that HyperFlow could use only one redis connection. Instead of using n blocking redis clients, we use single non-blocking command that is executed in loop - every checkInterval - and it resolves all relevant waiting getJobResult promises.
This PR changes they way of communication with remote jobs, that HyperFlow could use only one redis connection. Instead of using n blocking redis clients, we use single non-blocking command that is executed in loop - every
checkInterval
- and it resolves all relevant waitinggetJobResult
promises.With such changes
getJobResult
behaviour is identical, so there is no need to update functions or examples. However remote job executor must be adjusted accordingly - see https://github.com/hyperflow-wms/hyperflow-job-executor/pull/6.