Why can hosting slow down when disk space is available?
Storage, CPU, memory and concurrent tasks are different resources in a hosting account. Free disk space does not show that an application has enough CPU or RAM. 'Unmetered disk' also does not mean unlimited computing power or unlimited concurrent visitors.
This guide is for site owners who want to understand usage graphs in their panel and isolate a problem before upgrading. Verify exact limits and the accounts they apply to with your own service. The definitions below do not assign another provider's default quotas to Ajans Hosting packages.
What do CPU, RAM, inodes, I/O and EP measure?
| Indicator | Meaning | Do not confuse it with |
|---|---|---|
| CPU | CPU resources available for application processing. | It is not a visitor count or an exact page speed by itself; verify the unit displayed. |
| RAM / physical memory | The limit applied to memory used by processes running in the account. | It is not the same as the memory_limit setting for a single PHP process. |
| Disk space | The storage size of data within the account's scope. | File count and read/write speed are separate measures. |
| Inodes | A measure that tracks the number of filesystem objects, such as files and folders. | Many small files increase the count even when they use little space. |
| I/O | A limit on the amount of data read from and written to disk over time. | It is not network connection speed or a monthly traffic allowance. |
| IOPS | The number of disk read/write operations per second. | It is not the same as I/O; operation size changes the result. |
| EP / entry processes | A measure of tasks entering the resource-limiting environment simultaneously; implementation depends on the server. | It is not the total number of visitors present on the website at the same time. |
| NPROC | A limit on the total number of processes and threads in the account environment. | It does not measure the same scope as EP. |
CloudLinux's official limits documentation defines these resources separately and notes that EP implementation can differ between Apache and LiteSpeed environments. Panel labels, measurement methods and behaviour when a limit is reached may vary with your infrastructure.
Does a limit apply to a site, an account or the entire reseller service?
Multiple websites added to the same hosting account may share its resources. An import or backup job on one site can affect the resources available to others. The website count in a package table does not mean the same CPU/RAM amount is separately allocated to each site.
For reseller services, ask separately about subaccount limits and any overall reseller limits. '40 accounts' indicates the number of accounts that can be managed; it does not guarantee equal performance for forty busy stores. On a VPS, the operating system, database, panel and applications all use the virtual server's total resources together.
- Which account or system do the table's CPU/RAM values apply to?
- Which resources do multiple websites and subaccounts share?
- How are background jobs, email and databases included in measurements?
- Does exceeding a limit cause throttling, process errors or another policy action?
- Does the panel show an instantaneous value, an average or peak usage?
Confirm a limit breach within the same time interval
- Record the slow URL or operation and the incident time with its time zone. Distinguish website loading, administration screens, file transfers and checkout.
- Open Resource Usage if your panel provides it. Review CPU, RAM, I/O and process indicators for the incident time.
- Alongside average usage, look for limit violations or fault records if available. A brief spike may not appear in an average over a long interval.
- Check for backup, product import, cache generation or scheduled jobs running at the same time.
- After a change, measure the same operation again under similar conditions. Do not run load tests that put other websites or the live service under strain.
If the panel lacks the necessary detail, ask the provider for a summary of the relevant interval. Seeing '100%' on a graph does not mean the entire physical server is in use unless the unit and scope are known.
Look for evidence before attributing a symptom to resources
| Observation | Record to investigate | Initial assessment |
|---|---|---|
| CPU peaks during a particular operation | Application, query and task logs from the same time. | Is there an expensive query, plugin or repeated job? |
| 500 / 503 or a memory error | Application error logs and memory/process limit records. | An HTTP code alone does not prove insufficient RAM. |
| Slow performance during disk operations | I/O/IOPS and concurrent backup or transfer jobs. | Are jobs concentrated at the same times? |
| New files cannot be written | Disk space, inode count, quota and file permissions. | Distinguish space, object-count and permission issues. |
| 'Resource Limit Reached' appears | Error page, server type and relevant limit violation. | Match the exceeded limit against the provider's records. |
Example interpretation: Reaching a disk limit while a product import and backup run together calls for reviewing the task schedule first. This is an educational example; you need your own records to attribute your website's issue to the same cause.
Improve the workload first, then choose capacity
In WordPress, review unused plugins and repeated jobs, assess appropriate caching for public pages and schedule heavy imports at suitable times. The WordPress performance guide connects measurements with changes; in a store, also test the caching behaviour of personalised pages such as cart and checkout.
If inode usage is high, identify which areas produce the most files: caches, old backups, media variants and email archives require different approaches. Do not delete files in bulk without verifying a current backup and what the files do. Use the application's own clearing tool for caches and the business's retention plan for archives.
If normal, necessary workload still regularly reaches a limit after optimisation, consider a higher hosting package or a server model. If custom service permissions are also needed, use the hosting and VPS comparison. Remember that resource upgrades and application maintenance address different tasks.
Start the package discussion with a measurement summary
Service and account covered by the measurement: Affected website / operation: Incident time and time zone: CPU / RAM / I/O / process limits and units: Observed usage and any limit violation: Disk size and inode usage: Background jobs running at the same time: Fix attempted and remeasurement result: Expected growth or required custom service:
Do not attach complete logs containing secret keys or customer data to this template. For an existing service, request an investigation through the customer portal; for a new choice, compare hosting packages and get a quote based on your measurements.
Your next step
Review the service scope using the link below. For an account-specific action, contact the support center.
Compare hosting options →See all practical guides →
