"Resource Limit Is Reached" explained: CPU, I/O, entry processes and inodes on shared hosting
TL;DR: On shared hosting your account lives inside a fenced-off slice of a server. Systems such as CloudLinux LVE set caps for CPU, memory, disk throughput (I/O), disk operations (IOPS), total processes, simultaneous entry processes and, through disk quotas, the number of files (inodes). The "508 Resource Limit Is Reached" error usually means you hit the entry-process cap, meaning too many dynamic requests were running at once. Other limits behave differently: CPU and I/O slow your processes down, memory and process limits can kill them and an inode limit stops new files being written. Find which limit you hit in your panel's resource usage page, fix the cause (slow plugins, bots, backups, caching) and upgrade your plan only if the site has genuinely outgrown it.
What the error means
The message "508 Resource Limit Is Reached" appears when the web server cannot run another request for your account because a cap has been reached. The usual cause is the entry-process limit: your account was already running its maximum number of simultaneous dynamic requests, so new visitors were turned away.
The error does not mean the whole server is down. Other accounts on the same server are normally unaffected, because limits are applied per account. That is the point of the system: one noisy site should not slow everyone else.
Limits are a cause of the symptoms you may already know: slow pages in the admin area, intermittent 503 errors and sudden slowdowns during traffic bursts. See how many visitors shared hosting can handle for the arithmetic.
The limits, one by one
The CloudLinux documentation describes these limits. The defaults below are the recommended values for a typical shared account, but your host chooses its own numbers, so check your plan.
| Limit | Also called | Typical default | What it caps |
|---|---|---|---|
| CPU | SPEED | 100% (one core) | Processing time, as a share of one CPU core |
| Physical memory | PMEM | 1 GB | Real RAM your processes use, including shared memory and disk cache |
| Virtual memory | VMEM | 0 (unlimited, deprecated) | Address space across your processes |
| I/O | IO | 1,024 KB/s | Disk read and write throughput |
| IOPS | IOPS | 1,024 | Disk read and write operations per second |
| Processes | NPROC | 100 | Total processes and threads |
| Entry processes | EP | 20 | Concurrent dynamic requests and SSH or cron sessions |
| Inodes | Files | host-defined | Number of files and folders |
In plain terms:
- CPU is how much computing you can do.
- Memory is how much your running programs can hold.
- I/O and IOPS are how fast and how often you can read and write to disk.
- Processes are how many things can run at once.
- Entry processes are how many visitors' dynamic requests can be handled at the same instant.
- Inodes are how many files you can store.
What happens at each limit
The behaviour differs by limit, which helps you diagnose the problem.
| Limit hit | Typical result |
|---|---|
| CPU | Your processes are slowed down, not killed. Pages respond slowly |
| Memory | The system reclaims cache first, then kills processes in your account. You often see 500 or 503 errors |
| I/O | Processes are throttled and run slower |
| IOPS | Reads and writes pause until the next second |
| Processes | New processes cannot start, which can produce 500 or 503 errors |
| Entry processes | Requests are refused with the 508 Resource Limit Reached error |
| Inodes | New files cannot be written once a hard limit is reached |
This is why a site can crawl without any error message (CPU or I/O), or throw 508s even when its average CPU looks low (entry processes).
How to see what you hit
In cPanel on a host that uses CloudLinux, open Metrics, then Resource Usage. You will see graphs for CPU, memory, I/O, IOPS, processes and entry processes, with the account's limit, and a "Faults" count showing how often each limit was hit. A fault means the limit was hit at that moment.
- Open the page and choose the period, for example the last 24 hours.
- Look for faults. They show which limit is the problem.
- Note the times. Match them with your site logs to see what was running.
- Download the raw access log (cPanel's Raw Access) and look for spikes, repeated requests or heavy URLs.
If your host does not show a page like this, ask support which limit your account is hitting. Our guide to server monitoring beyond uptime explains useful metrics on a VPS.
Common causes
- Traffic spikes. A sale, newsletter or social mention.
- Bots and scrapers. Crawlers, vulnerability scanners and AI crawlers that make thousands of dynamic requests. See blocking AI crawlers.
- Brute-force login attempts against
wp-login.phpandxmlrpc.php. - Slow pages. Each slow request occupies an entry process longer, so fewer fit at once.
- Admin activity. The WordPress dashboard and Heartbeat API generate steady dynamic requests. See fixing admin-ajax.php high CPU.
- Backups, imports and scans running during busy periods.
- Cron jobs that overlap or run too often. See WordPress cron jobs.
- No page caching. Every visit runs PHP.
- Heavy plugins and themes.
- Large databases or slow queries.
- Mail loops or runaway scripts.
How to fix each limit
Entry processes (508 errors)
- Turn on page caching so anonymous visitors do not run PHP.
- Make slow pages faster: fewer plugins, optimised database, current PHP.
- Block or rate-limit bots and brute-force traffic at the firewall or CDN.
- Slow Heartbeat down and close idle dashboard tabs.
- Move heavy tasks, such as backups and imports, to quiet hours.
CPU
- Cache pages and objects.
- Remove or replace plugins with high CPU use. Query Monitor helps find them.
- Update PHP. See PHP 8.2 end of life and how to upgrade.
- Use a CDN to offload static content. See do you need a CDN.
Memory
- Reduce plugins, and raise the PHP memory limit only if your plan allows. See PHP memory limit and execution time and the memory exhausted error.
- Avoid running big imports or image processing during the day.
I/O and IOPS
- Reduce disk-heavy tasks: large backups, log growth and big file scans.
- Use caching so fewer database and file reads happen.
- Compress and optimise images in advance, not on the fly.
Processes (NPROC)
- Check for stuck or runaway scripts, and for cron jobs that spawn too many processes.
- Review plugins that start many background processes.
Inodes: the file-count limit
Inodes count every file and folder in your account, including emails, cache files, backups, logs and image thumbnails. The limit is enforced through the disk quota on the home partition, so it behaves differently from the other limits: once a hard limit is reached, new files cannot be written, and uploads, cache creation, email delivery and backups can fail. A soft limit may allow a grace period, which defaults to seven days in the CloudLinux documentation.
Common causes of inode problems:
- thousands of cache files,
- old backups stored in the account,
- email accounts holding many small messages,
- generated image sizes from many plugins and themes,
- log files and session files that are never cleaned up.
How to reduce inodes:
- Use the cPanel Disk Usage tool to find folders with the most files.
- Clear cache folders and remove old backups. Keep backups off-site. See how to back up a WordPress site and test the restore.
- Delete unused plugins, themes and staging copies.
- Clean up old email and set auto-delete for spam and trash.
- Remove unneeded image sizes and regenerate thumbnails.
When to ask your host to raise limits
Ask politely if:
- you have fixed obvious inefficiencies and still hit limits during normal traffic,
- you have a one-off event such as a launch or sale and know the dates,
- you need a short-term increase while you migrate.
Give your host the time of the faults, the limit hit, what you have already done and the plan you expect. Hosts are more likely to help with evidence than with a general complaint.
Note that shared hosts usually cannot raise limits beyond what the plan allows. For sustained growth, a larger plan is the answer.
When to upgrade
Consider a move up when:
- faults appear repeatedly at normal traffic levels,
- your site needs memory or CPU the plan cannot provide,
- you run a store or membership site where caching helps little,
- the time you spend firefighting costs more than the price difference.
Options include a higher shared or managed tier, a cloud plan or a VPS. See cloud hosting versus VPS and how to choose a VPS plan in India.
Not every host works this way
CloudLinux is common, but not universal. Other hosts may use different tools, different names and different defaults. Some give you dedicated resources on cloud or VPS plans, where limits are the machine's own CPU, RAM and disk, and where you manage them yourself. Check your host's documentation for how it measures and enforces limits.
FAQs
What does "508 Resource Limit Is Reached" mean?
Your hosting account hit a cap set by the host, most often the limit on simultaneous dynamic requests (entry processes), so the server refused new requests.
Is the whole server down when I see this error?
Usually not. Limits are applied per account, so other accounts are normally unaffected.
What is an entry process?
A concurrent dynamic request, such as a PHP page being generated, or an SSH or cron session, counted against your account.
How do I check which limit I hit?
In cPanel with CloudLinux, open Metrics, then Resource Usage and look at the faults for each limit. Otherwise ask your host.
What are inodes?
The number of files and folders in your account. When the limit is reached, you can no longer create new files.
Will caching fix it?
Often, yes for anonymous traffic, because cached pages do not use PHP. It does not help much with logged-in users, carts or checkout.
Can my host raise my limits?
Sometimes, for a short period, if you can show the cause. For sustained needs, move to a bigger plan.
Do I need a VPS to avoid these errors?
Not always. Fix inefficiencies and add caching first. Move to a VPS when your site genuinely needs more resources than shared hosting provides.
Conclusion
A resource-limit error is your host telling you which cap your account hit. Read the faults to learn which one it was, fix the cause with caching, bot control, lighter plugins and cleaner storage, and upgrade only when the site has really outgrown its plan.
HostCloud makes plan limits clear and gives you upgrade paths from shared to VPS as you grow. See the plans at https://hostcloud.in.
Sources: