Specialisation rules

Routing particular jobs to particular managers. The tool for THE REPORT THAT BLOCKS EVERYTHING ELSE, and THE WAY IT IS USUALLY MISCONFIGURED. The problem specialisation solves is concrete. One report runs for six hours. It occupies a standard manager process for six hours. Two or three of those and the standard manager is full, so every short request in the estate queues behind them — and a user waiting twenty minutes for a two-second report is looking at a system that appears broken. THE FIX IS A DEDICATED MANAGER FOR THE LONG JOBS, AND EXCLUDING THEM FROM THE STANDARD ONE. The long reports run in their own lane with their own capacity; the standard manager stays free for everything else. Nothing runs faster; the queue simply stops being blocked by work whose duration is expected. AND THE USUAL MISCONFIGURATION IS DOING ONLY HALF OF THAT. A dedicated manager is created…

In this module: Module 8 · Concurrent processing

1. 1What a concurrent request is
2. 2Managers, queues and workers
3. 3Specialisation rules
4. 4Work shifts
5. 5A request that will never finishFree preview
6. 6Reading a request log
7. 7Killing a request safely
8. 8Managers that will not start
9. 9Output, printing and delivery
10. 10Purging: the table that ate the database
11. 11What breaks: four concurrent processing failures