508 Resource Limit Is Reached (cPanel / CloudLinux LVE)

What’s actually happening

The site throws '508 Resource Limit Is Reached' under load and recovers on its own once traffic drops. It tends to hit at peak hours or during a traffic spike, then clear minutes later. Static pages may still load while PHP-heavy ones fail, because the cage caps concurrent entry processes separately from raw CPU. In WHM the account shows faults piling up against one of its LVE limits.

Common causes

• The account hit the entry-process (EP) limit — CloudLinux defaults this to 20 concurrent entering processes, and a traffic burst or slow PHP fills all of them.
• CPU or physical-memory (PMEM) limit for the cage was reached, commonly from an unbounded query, a runaway plugin, or an inefficient theme.
• I/O or IOPS throttling kicked in from heavy disk activity — large backups, log writes, or a database doing full scans.
• Bad bots, aggressive crawlers, or a brute-force run against wp-login.php / xmlrpc.php spawning many simultaneous PHP processes.
• The plan's limits are simply too low for the site's real concurrency — fine until it grew.

How to fix it

1. Read the LVE fault history firstIn WHM go to CloudLinux LVE Manager and check which resource is faulting for the account — EP, CPU, PMEM, or IO. Fixing blind is wasted effort; the four causes need different fixes. lveinfo --by-fault from the command line gives the same data.
2. Raise the limit that's actually being hitWHM > CloudLinux LVE Manager > edit the account and bump the specific limit. If EP is the wall, the default 20 is often too low for a busy site — raising it to 30-50 (and NPROC to ~100) is typical. Don't inflate everything blindly; raise what the fault log named.
3. Cut the load at the sourceAdd a page cache (LiteSpeed Cache, WP Super Cache) so most hits never invoke PHP. Block abusive IPs and bots, and protect wp-login.php / disable xmlrpc.php if a brute-force is driving the process count. This often fixes EP exhaustion without touching limits.
4. Find the expensive code or queryIf CPU/PMEM is faulting, enable the slow query log and PHP profiling to find the offending plugin, cron job, or query. A single unindexed query run on every request can hold processes open long enough to exhaust the cage.
5. Upgrade the plan if the site has outgrown itWhen limits are correctly tuned and load is legitimate, the account genuinely needs more CPU/RAM/EP than the plan allows. Move to a higher tier or off shared hosting. No amount of tuning beats real demand exceeding the cage.