
ALT Simulation of three workloads for the simulated queue. This screenshot was taken under different simulation conditions to the previous one. The Interactive User fires a job every 5 seconds and measure latency. The Controlled Loads adjust their firing rate to maintain target workloads and average latency measurements over a minute. Their jobs are of random length up to 30s. Here the high priority work is impacting the system's ability to run the low priority work.

ALT Simulation dashboard showing the simulated platform queue. This implementation has a simple "Run priority jobs first" implementation, so allows low priority jobs to be starved out. It has been set with few parallel executors in order to force contention, and it penalises all jobs with a random startup delay averaging 76ms at the time this screenshot was taken.




















