Understanding Starvation

[LRagji]

This is a good library and provides great abstraction to underlying worker methods so thanks for the effort, recently i was monitoring my application with the properties that have been provided on the class and i am seeing a strange pattern can someone explain what does it mean ..

Total Time required by Async Method : 6086ms
Total Threads : 9
Utilisation: 69%
Waiting Time: 4208ms
Runtime Average: 604ms
and this repeats .....

• Is my application starving on threads?
• Why is there a huge wait time when runtime time is so small?
• If everything is good then why is the method taking 6 second instead of the actual runtime + some tolerance?

Config idleTimeout:3000; concurrentTasksPerWorker:1

[jasnell]

The total time includes waiting time, runtime, and the messaging passing overhead which depends quite a bit on the timing of the event loop. If you enable event loop delay monitoring it might give you an idea of how much latency event loop delay is introducing.

[LRagji]

Which would explain the “total time” composition,
What still seems to be missing is why is there a wait time when the utilisation is just 69% ? Should it not be 100% for wait time to be more than 0?