Concurrency

Configuration

Control how many profiles run in parallel:


const params = {
  execution: {
    concurrency: 3,               // 3 profiles at once
    delayBetweenProfilesMs: 2000, // 2s delay between starts
  },
}

Example 1 — Sequential execution


const params = {
  execution: {
    concurrency: 1,  // one profile at a time
  },
}

Example 2 — High concurrency with delays


const params = {
  execution: {
    concurrency: 10,              // 10 parallel profiles
    delayBetweenProfilesMs: 3000, // stagger starts by 3s
  },
}

Log output during delay:


⏳ Waiting 3000ms before next profile...

How it works

The SDK uses p-limit to control concurrency:


Timeline (concurrency: 3, delay: 1s):

  t=0s   Profile 1 starts ─────────────── done
  t=1s   Profile 2 starts ──────────── done
  t=2s   Profile 3 starts ───────────────── done
  t=3s   Profile 4 starts (slot freed) ──── done
  ...

Window Manager

The SDK includes WindowManager to calculate grid positions for browser windows:


import { WindowManager } from '@hira-core/sdk'
 
const screen = { width: 1920, height: 1080 }
 
const pos = WindowManager.calculatePosition(
  0,     // profile index
  screen,
  800,   // window width
  600,   // window height
  0.5,   // scale factor
)
// → { x: 0, y: 0 }
 
const pos2 = WindowManager.calculatePosition(1, screen, 800, 600, 0.5)
// → { x: 800, y: 0 }

Example 1 — Default grid layout

With 1920×1080 screen and 800×600 windows:


┌──────────┬──────────┐
│ Profile 0 │ Profile 1 │
├──────────┼──────────┤
│ Profile 2 │ Profile 3 │
└──────────┴──────────┘

Example 2 — Scaled windows for more profiles


// 50% scale = 400×300 per window → fits more profiles on screen
const pos = WindowManager.calculatePosition(index, screen, 800, 600, 0.5)

Performance Tips

Start small — test with concurrency: 1 first
Add delays — delayBetweenProfilesMs prevents rate limiting
Monitor memory — each browser profile uses ~200-500MB RAM
Scale gradually — increase concurrency based on system resources