Run a long running task in Kotlin (Ktor backend) - kotlin

Question

In my Ktor backend, I have a simple get API request. It will make some database update and finally return a message. But, before returning the response, the API will initiate a long-running task (some database operation and file writing tasks) which might take some time to finish. The long-running task should continue its job even if the API returns its response.
This is my get API request
routing {
get("/testAPI") {
val x= queryResolver()
println(" return from API")
call.respondText("Hello ..")
}
}
Below are my handling functions. queryResolver() is called from the API and then it will call myService() which will call db_update1(), db_update2() and finally starts the long_running_task() .
suspend fun queryResolver(): String{
println(" queryResolver- start -${Thread.currentThread().name}")
delay(2000)
val x = myService()
println(" queryResolver - end")
return "..return from queryResolver "
}
suspend fun myService(): String{
println(" service - start -${Thread.currentThread().name}")
delay(2000)
val x = db_update1()
val y = db_update2()
GlobalScope.launch() {
println(".. launch start -${Thread.currentThread().name} ")
long_running_task()
println(".. launch - end ")
}
println(" service - end")
return "..return from service -> "
}
suspend fun db_update1(): String{
println(" db_update1 - start -${Thread.currentThread().name}")
delay(2000)
println(" db_update1 - end")
return "..return from db_update1 -> "
}
suspend fun db_update2(): String{
println(" db_update2 - start -${Thread.currentThread().name}")
delay(3000)
println(" db_update2 - end")
return "..return from db_update2 -> "
}
suspend fun long_running_task(): String{
println("### long_running_task - start -${Thread.currentThread().name}")
delay(10000)
println("### long_running_task - end")
return "..return from long_running_task -> "
}
This code gives the below output which is desired output.
Start API
queryResolver- start -eventLoopGroupProxy-4-2
service - start -eventLoopGroupProxy-4-2
db_update1 - start -eventLoopGroupProxy-4-2
db_update1 - end
db_update2 - start -eventLoopGroupProxy-4-2
db_update2 - end
service - end
queryResolver - end
return from API
.. launch start -DefaultDispatcher-worker-1
### long_running_task - start -DefaultDispatcher-worker-1
### long_running_task - end
.. launch - end
But, using the global scope is not a good idea in this case.
Now, what is the best way to keep alive(to allow it to finish its task) the long_running_task() even if the API response returns?
If I want to reuse the scope and the context of the existing scope and context how can I do that?