Shared Hosting CPU Limits Explained (and Why Your Site Keeps Crashing)

1. "Unlimited" Is a Lie — Here's the Fine Print

Every major shared host advertises "unlimited" disk space, "unlimited" bandwidth, "unlimited" email accounts. Then read the acceptable-use policy: unlimited doesn't mean unlimited. It means "subject to our resource limits, which we reserve the right to enforce at our discretion."

What they're really selling is bounded access to a shared machine. Almost every cheap host runs CloudLinux OS, which uses Lightweight Virtual Environments (LVE) to give each hosting account a hard cap on CPU, RAM, I/O, and concurrent processes. The moment you cross any cap, your site is throttled or requests are dropped.

This is not a scam — it's how you can rent part of a server for $3/mo. But the marketing pretends the limits don't exist, which is why so many first-time site owners are blindsided when their modestly-trafficked blog starts returning 508 errors after a Reddit mention.

2. The Real Resource Limits, Decoded

There are seven resource limits that matter. Each cheap host sets them slightly differently, but the ranges are similar. These are the values we've seen on entry-level shared plans at Bluehost, Hostinger, HostGator, GoDaddy, Namecheap, and A2:

The one that kills most sites: EP (Entry Processes)

EP is the single most common cause of "sudden outage" on shared hosting. An Entry Process is one active dynamic request — basically one PHP-FPM worker handling one page load. An EP limit of 20 means only 20 visitors can be simultaneously loading a dynamic page. Visitor 21 gets a 508 error.

If your pages take 2 seconds to load and your EP limit is 20, you can handle 10 visitors per second peak — about 864,000 visits/day in theory, but real traffic is bursty and your peak-to-average ratio is usually 5–10×. So a $3/mo plan with 20 EP can realistically serve only ~100k visits/day before hitting 508s on spikes.

I/O and IOPS — the silent speed killers

An IOPS limit of 1,024 means 1,024 disk operations per second. WordPress fires roughly 50–200 disk operations per page load. At 1,024 IOPS you're capped around 5–20 page loads per second, per account. This is why so many "fast enough" plugin stacks suddenly freeze when a second visitor arrives.

3. How to Measure Your Own Limits

If your host uses cPanel

Log into cPanel and look for "Resource Usage" or "CPU and Concurrent Connection Usage". CloudLinux surfaces a graph showing when you hit each limit in the last 24 hours, 7 days, or 30 days. If you see red bars, you're being throttled.

If you have shell access

lvectl list-user $(whoami)

Outputs your exact CPU, PMEM, EP, NPROC, and IO limits. This is the source of truth — what's advertised isn't always what's provisioned.

If you only have PHP

Drop this into a PHP file:

<?php
echo "Memory limit: " . ini_get('memory_limit') . "\n";
echo "Max exec time: " . ini_get('max_execution_time') . "\n";
echo "Max input vars: " . ini_get('max_input_vars') . "\n";
echo "Upload max: " . ini_get('upload_max_filesize') . "\n";
?>

A memory_limit of 128M on a "Business" plan tells you exactly how much headroom WordPress has before it crashes.

Synthetic load test

Use Loader.io or k6 to simulate 50, 100, 200 concurrent users. Watch for the point where error rate jumps above zero — that's where your EP limit lives. Do this on a staging site, not production.

4. What Happens When You Hit a Limit

Different limits fail differently, and the error message usually points at the cause — if you know what you're looking at.

• CPU throttling — requests don't fail, they slow down. TTFB doubles or triples. Users notice, bounce rate climbs.
• EP limit hit — HTTP 508 "Resource Limit Reached" or HTTP 503 "Service Unavailable". New visitors get an error page. Cached pages still load.
• NPROC limit — cron jobs fail silently, SSH sessions get killed, WP-Cron stops running. The site looks fine but backups, scheduled posts, and WooCommerce emails quietly stop.
• Memory limit — PHP fatal: "Allowed memory size of X bytes exhausted". A white page, or just a broken admin dashboard.
• IOPS throttling — everything slows, including the cPanel itself. File-manager operations stall.
• Inode limit — cannot write new files. Email delivery breaks, new uploads fail, backup plugins choke.

5. How to Fix Resource Errors

Short-term fixes (free)

1. Install page caching. LiteSpeed Cache or W3 Total Cache. A page served from cache costs ~0 EP and ~0 IOPS. This alone can 10× your real capacity.
2. Put Cloudflare in front. Cloudflare caches HTML at the edge (especially with APO), so most requests never reach your host.
3. Audit plugins. Use Query Monitor to find the slowest queries. Remove anything firing 100+ queries per pageview.
4. Move WP-Cron off real-time. Disable WP-Cron in wp-config and add a real system cron running every 5 minutes — eliminates admin-ajax spam.

Medium-term fixes ($)

1. Upgrade to the host's "Business" or "Pro" plan — usually bumps EP from 20 to 40 and CPU from 1 core to 2.
2. Enable object caching (Redis or Memcached) to slash IOPS.

Long-term fix (right answer)

Move to hosting with baseline limits high enough that you never think about them. Our Growth plan gives you 4 cores, 4 GB RAM, 100 EP, 50 MB/s I/O, and 8,192 IOPS on NVMe — 5–10× the headroom of a typical $3/mo plan, for $5.99/mo. Or jump straight to a VPS for dedicated cores and RAM with no neighbours.