Hosted browser-grid SaaS platforms bill $150–$600/mo per parallel WebDriver slot. Using BotFoundry’s strict 1 dedicated thread + 1.5 GB RAM per active session rule with shm_size: 2gb and automatic 50-session container draining, teams run 10 to 40+ parallel Chrome/Firefox sessions on flat-rate KVM or Ryzen hardware.
Public rates retrieved 2026-11-12 from official vendor pricing pages cited below; promotional & renewal prices may change — always verify exact specs, term length, and renewal rates on the vendor’s official checkout page.
| Concurrency Scale | vCPU | DDR5 RAM | OS / Runtime | /dev/shm & Watchdog Guard | Public Tier Match (2026-11-12) |
|---|---|---|---|---|---|
| 4 Parallel WebDriver Slots | 4 vCPU KVM | 8 GB DDR5 | Ubuntu 24.04 Docker Grid | shm_size: 2gb + Timeout 180s |
Hostinger KVM 2 ($5.99/mo) |
| 10 Parallel WebDriver Slots | 8 vCPU KVM | 16–32 GB DDR5 | Ubuntu 24.04 Docker Grid | Drain=50 + Localhost 4444 |
Hostinger KVM 8 ($19.99/mo) |
| 20 Parallel Slots + Video Sidecar | 8–12 Dedicated VDS | 32 GB ECC | Ubuntu 24.04 Dedicated | Dedicated Threads + NVMe |
UltaHost VDS 32GB / InterServer 32GB |
| 40+ Multi-Browser Crawler Fleet | 12–16 Core Ryzen | 64–128 GB DDR5 | Bare-Metal Ubuntu 24.04 | Bare-Metal IPC + Price-Lock |
InterServer Ryzen 64GB ($89/mo) |
services:
selenium-hub:
image: selenium/hub:4.26.0
restart: unless-stopped
ports:
- "127.0.0.1:4444:4444"
environment:
SE_SESSION_REQUEST_TIMEOUT: 120
SE_NODE_SESSION_TIMEOUT: 180
chrome-node:
image: selenium/node-chrome:4.26.0
restart: always
init: true
shm_size: 2gb
depends_on: [selenium-hub]
deploy:
replicas: 4
resources:
limits:
memory: 3500M
environment:
SE_EVENT_BUS_HOST: selenium-hub
SE_EVENT_BUS_PUBLISH_PORT: 4442
SE_EVENT_BUS_SUBSCRIBE_PORT: 4443
SE_NODE_MAX_SESSIONS: 2
SE_DRAIN_AFTER_SESSION_COUNT: 50Long-lived ChromeDriver and Xvfb virtual framebuffer processes leak file descriptors across hundreds of E2E runs. Setting SE_DRAIN_AFTER_SESSION_COUNT=50 alongside restart: always gracefully drains each node after 50 sessions and boots a clean container.
Stuffing 8 Chrome sessions onto a 2-vCPU VM starves DOM layout threads and triggers flaky TimeoutException failures on element waits. Budget 1 vCPU thread and 1.5 GB RAM per concurrent browser session (plus +0.35 vCPU if FFmpeg video recording sidecars are enabled).
An unauthenticated Selenium Grid Hub port 4444 exposed on a public IP is routinely compromised within minutes by automated scanners launching arbitrary Chrome processes — always bind 127.0.0.1:4444:4444 or guard via private VPN.
Enabling selenium/video FFmpeg sidecars adds ~0.35 vCPU and continuous sequential NVMe writes per active browser — step up to Dedicated VDS or InterServer Ryzen hardware when recording all runs.
Enforce SE_NODE_SESSION_TIMEOUT=180 (3-minute idle session reaper), SE_SESSION_REQUEST_TIMEOUT=120, and pageLoadStrategy: "eager" inside your WebDriver client options.
Once concurrency crosses 20–25 heavy Chrome/Firefox sessions running 24/7, a bare-metal 12–16 core Ryzen node with 64 GB DDR5 and flat port quota outperforms stacked virtual machines on price and IPC.
BotFoundry 4-layer RAM equation, /portable chart trimming, and Windows AutoLogon watchdog rules for 1 to 50+ MetaTrader 4/5 Expert Advisor terminals.
BotFoundry sizing formulas for Ubuntu 24.04 Freqtrade Docker bots, cgroup MemoryMax bounds, chronyd NTP sync, and isolated FreqAI LightGBM retraining.
BotFoundry memory math for headless Chromium BrowserContexts, Docker --shm-size=2g shared memory pools, tini zombie reaping, and flat-bandwidth KVM nodes.
BotFoundry 4-layer RAM guide for scaling self-hosted n8n Queue Mode with dedicated webhook receivers, Redis AOF queues, and PostgreSQL 16 pruning.