A hosting performance audit helps you separate server problems from application code, databases, networks, and third-party services. This repeatable workflow shows what to measure, how to test fairly, and when to escalate, optimize, or compare providers.
Overview
A slow website does not automatically mean your hosting provider is at fault. A page can be delayed by an overloaded database, an unoptimized image, a blocked JavaScript request, a slow DNS response, a plugin, or an external service. Conversely, application-level optimization will not fix a server that is consistently short of CPU, memory, or database capacity.
The purpose of a hosting audit is to establish a baseline and then compare results under consistent conditions. Record the date, URL, test location, device type, browser or testing tool, cache state, and any recent changes. Without that context, two speed tests can appear to contradict each other even when both are accurate.
Use the audit for a single site or for a fair hosting comparison. The goal is not to chase one unusually fast result. Look for patterns across several tests and over time.
What to track
1. Page experience and server response
Test representative pages rather than the homepage alone. Include a lightweight page, a content-heavy page, a search or category page, and—if applicable—a logged-in or checkout flow. Track:
- Largest Contentful Paint (LCP): how quickly the main visible content becomes available.
- Interaction to Next Paint (INP): how quickly the page responds to user interactions.
- Cumulative Layout Shift (CLS): whether visible content moves unexpectedly during loading.
- Time to first byte (TTFB): the delay before the server begins returning a response.
- Fully loaded time and request count: useful diagnostic context, though they should not replace user-focused metrics.
Use a browser-based test and, where possible, field data from real visitors. Run tests with a cold cache and a warm cache because visitors may encounter either condition. Mobile and desktop results should be recorded separately.
2. Uptime and availability
Use an external monitoring service to check the site at regular intervals from more than one location when possible. Record the timestamp, response code, response time, outage duration, and whether the failure affected the entire site or only a feature. A monitoring alert is evidence of an event, not always proof of its cause: DNS, certificates, networks, application errors, and hosting infrastructure can all produce different symptoms.
Include HTTPS and, if important to the business, a transaction or login check. A homepage check may succeed while a database-backed workflow is failing.
3. Server and account resources
Review the resource panel or logs supplied by your host. The exact labels vary, but useful signals include CPU usage, memory usage, process or entry limits, disk space, inode counts, disk input/output, database connections, and bandwidth. Note both current usage and the period during which limits were reached.
For WordPress and other PHP applications, also record the PHP version, worker or process limits, slow requests, scheduled tasks, and database size. A short CPU spike is less concerning than a recurring limit that coincides with slow responses or errors. This resource-limit guide provides useful context when deciding whether to optimize or upgrade.
4. Dependencies outside the host
List content delivery networks, DNS providers, payment tools, fonts, analytics, advertising scripts, APIs, image services, and email systems. In a waterfall view, identify requests that are slow, blocked, or repeatedly failing. A third-party script can affect the visible page even when the origin server responds quickly.
A simple evidence log keeps the audit repeatable:
| Date and time | URL and test type | Location/device | Key result | Recent change or symptom | Next action |
|---|---|---|---|---|---|
| Record consistently | Homepage, article, checkout, or API | Document every test condition | TTFB, LCP, INP, CLS, uptime, or error | Plugin, deploy, traffic spike, or no change | Retest, optimize, or contact support |
Cadence and checkpoints
Run a lightweight review monthly and a fuller review quarterly. Monthly checks should cover uptime alerts, resource-limit notifications, representative page tests, certificate validity, and recent support tickets. A quarterly review should examine trends, database growth, traffic patterns, backups, software versions, and whether the current hosting plan still matches the site’s workload.
Repeat the baseline after meaningful changes, including a theme or plugin update, a framework or PHP upgrade, a DNS or CDN change, a migration, a caching configuration change, or a major content release. Test before and after the change when practical. Keep the same URLs and test conditions so the comparison is useful.
For a provider comparison, create identical test sites or compare sites with closely matched software and content. Document location, cache state, plan type, server region, and configuration. A single run is not a hosting speed test with enough context to support a purchasing decision. For broader trade-offs between shared hosting, VPS, and cloud options, see this hosting model comparison.
How to interpret changes
Start with timing. If performance worsened immediately after a code, plugin, database, or configuration change, investigate that change first. If several unrelated pages slow at the same time while resource limits or server errors rise, the host or hosting plan becomes a stronger suspect. If only one page is affected, inspect its queries, media, scripts, and content.
- High TTFB with a small page: investigate server processing, database queries, PHP workers, caching, DNS, and network distance.
- Normal TTFB but slow visual loading: inspect images, render-blocking assets, fonts, JavaScript, and third-party requests.
- Slow only for uncached requests: review origin performance, cache rules, object caching, and database work.
- Frequent resource-limit events: identify the process causing the usage, reduce unnecessary work, or evaluate a larger plan.
- Intermittent outages from one location: compare external monitors before concluding that the origin server is down.
- Errors during traffic peaks: correlate request volume, application logs, database connections, and account limits.
Do not treat a high score from one lab tool as a complete diagnosis. Use the waterfall, server logs, real-user measurements, and resource history together. When contacting support, provide timestamps, affected URLs, request IDs if available, monitoring results, and a concise description of what changed. This gives the provider something actionable to investigate.
When to revisit
Schedule the next audit on a monthly or quarterly cadence, depending on the site’s traffic and business importance. Revisit immediately after an outage, traffic spike, migration, deployment, caching change, domain or DNS update, or noticeable decline in search or conversion performance. Also review the evidence before renewing a hosting plan or comparing alternatives.
At each review, preserve the previous baseline rather than overwriting it. Compare medians and recurring patterns, not just the best or worst individual result. Mark unresolved issues, assign an owner, and set a retest date. If the host cannot explain repeated resource exhaustion or availability events, collect the evidence and assess whether a plan change or migration is justified. Before moving, follow a documented website migration checklist, then repeat the same tests on the new environment.
Practical next step: choose three representative URLs today, create an evidence log, run tests under documented conditions, and save the results. Repeat the same checks in 30 days. That simple baseline will make future optimization, support conversations, and hosting comparisons more reliable.