Lightweight Virtual Environment (LVE) is a kernel-level technology developed by CloudLinux, used to control and isolate resource usage on shared hosting servers.
The primary purpose of LVE is to ensure that no single hosting account can negatively impact overall server performance by consuming excessive system resources.
Each account is allocated defined limits for CPU, memory, processes, and disk I/O. Information defining these limits can be found below.
NOTE:
Though this is to ensure no single site affects the overall server performance, setting these too high, or having multiple accounts allowed to use more than the overall servers available resources can have issues.
Whenever sites hit any of these defined limits, the site will usually display a ‘508 Resource Limit Is Reached’ to ensure sites arent using more than their limits.
CPU (SPEED)
CPU limits define how much processing power an account can consume. LVE CPU usage is measured as a percentage of a single CPU core, not the entire server.
For example:
A limit of 100% CPU equals one full CPU core
On a quad-core server, this equates to 25% of total server CPU capacity
When an account reaches its CPU limit:
Existing processes continue running
New processes are throttled (put to sleep) until resources are available
This commonly results in:
Slower page loads
Delayed responses
Timeouts during high activity
NOTE:
CloudLinux strongly recommend setting CPU speed limits no lower than 100%.
Such limits cause CPU context switching which leads to increased%sys.
Physical Memory (PMEM)
Physical Memory is the actual RAM actively in use by the account’s processes.
When an account exceeds its physical memory limit:
CloudLinux immediately restricts further memory usage
The website will typically display a 508 Resource Limit Reached error
These errors are usually temporary, and automatically clear once memory usage drops below the defined limit.
Virtual Memory (VMEM)
Virtual Memory represents the maximum amount of memory that all processes within an LVE are allowed to allocate, including memory that may be swapped to disk.
When a process attempts to allocate memory:
CloudLinux checks whether the total virtual memory usage remains within the assigned limit
If the limit would be exceeded, the allocation is denied.
In most cases, hitting the VMEM limit causes the affected process to fail or terminate, which can result in application errors or blank pages.
NOTE:
It is recommended to disable VMEM limits (set them to 0) in your system at all because they are deprecated in CloudLinux OS 6 and 7 system, and can cause unexpected issues.
Entry Processes (EP)
The Entry Processes limit defines how many concurrent requests an account can handle at a given moment.
Common entry processes include:
PHP page executions
SSH sessions
Cron jobs
A frequent misconception is that EP represents the number of visitors on a site.
While each PHP request does create an entry process, these processes typically execute, run and end very quickly.
However, a large volume of EP may spawn if you have a high influx of simultaneous visitors.
Reaching the EP limit usually indicates:
A sudden spike in traffic
Long-running PHP scripts
Poorly optimised applications or plugins
Number of Processes (NPROC)
This limit controls the total number of processes an account can run simultaneously, including all background and system tasks.
Unlike EP, this includes:
All PHP processes
Shell commands
Background jobs
This limit is rarely reached under normal conditions, as most non-PHP processes execute and terminate very quickly.
I/O Usage (IO)
I/O limits control how much disk activity an account can generate, and is measured in KB/s.
When an IO limit is reached, new processes are throttled (put to sleep) until resources are available.
Any operation that reads from or writes to disk counts toward I/O usage, including:
Database queries
File uploads and downloads
Log writing
Cache generation
I/O limits prevent a single account from saturating disk performance, which would otherwise degrade performance for all users on the server.
IOPS
IOPS limits restrict the total number of read/write operations per second.
When an IOPS limit is reached, user read/write operations stop until the current second expires.
NOTE:
More info on LVE Limits can be found here: