Hello, Jenkins geeks.
Once upon a time, I had a bunch of the downstream jobs to run in parallel. But, the number was quite big… Let’s limit parallel builds!
This is easy!
import java.util.stream.Collectors
...
Map parallelResults = [:]
List parallelData = [
"service1", "service2", "service3", "service4", "service5", "service6", "service7", "service8", "service9", "service10", "service11", "service12", "service13", "service14", "service15", "service16",
]
if (parallelData.size() == 0) {
log.info('Looks like there is nothing to build...')
} else {
def parallelCount = 5
def parallelSkip = 0
def parallelTotal = parallelData.size()
def parallelLimitReached = 0
while (parallelTotal != parallelLimitReached) {
List parallelDataSlice = parallelData.stream().skip(parallelSkip).limit(parallelCount).collect(Collectors.toList())
parallelLimitReached += parallelDataSlice.size()
parallelSkip += parallelCount
Map parallelSteps = parallelDataSlice.collectEntries {
return ["Building ${it}" : getParallelStep(it)]
}
parallelSteps.put('failFast', false)
parallelResults.putAll(parallel(parallelSteps))
}
}
...
def getParallelStep(String serviceName) {
return {
build(
job: serviceName,
parameters: [
string(name: "parameter", value: "value"),
],
propagate: false,
wait: true
)
}
}
Parameter wait is mandatory!
So, concurrency is 5 builds at a time (parameter parallelCount). The only issue here is that all 5 should finish job run, only after that next 5 will be queued up.
Happy Scripting!