How to Improve Drupal Performance: Find the Bottleneck Before You Buy a Fix

Light

post-banner
You’ve enabled caching, cleared it and checked again, but your Drupal site is still slow. Now what: upgrade hosting, optimize the database, change the code or add more caching? The answer depends on where the bottleneck sits. That’s because Drupal performance depends on factors across different layers of the stack, with each layer introducing different potential causes of slowdowns. 
To improve Drupal performance, the first step is to find out where the slowdown comes from. Without that diagnosis, it’s easy to spend time and money fixing the wrong problem. 

 

 

What to Know Before You Invest in Performance Fixes 

  • Most Drupal performance problems develop gradually rather than appearing overnight. 
  • Performance issues that look identical to users often have completely different underlying causes. 
  • The most effective optimization efforts begin with identifying the layer responsible for the slowdown. 
  • Early diagnosis helps teams focus on effort where it will have the greatest impact and avoid solving the wrong problem. 
  • Large investments in infrastructure or redevelopment are often unnecessary until the root cause is understood. 

 

 

Why Drupal Sites Slow Down 

Drupal site performance often slows down gradually as functionality and configuration accumulate, making the underlying cause harder to identify. 
As a site evolves, modules, fields, views and content types add to the database and increase Drupal’s rendering work. Over time, some of the changes may no longer serve a clear purpose or reflect how the site is used today. The current team may inherit years of technical decisions whose original rationale is no longer clear, making it harder to identify what is contributing to performance degradation. 
Drupal has reduced some performance overhead in newer versions, but older sites don’t benefit from that shift. Drupal 11.3 reduced database queries and cache operations on cold-cache requests by about a third, its biggest performance gain since Drupal 8. But upgrading alone won’t remove years of accumulated configuration and functionality. 
Similarly, turning on caching or changing a setting may not solve the problem. When performance issues stem from dozens of small changes, there is rarely a single fix that reverses them. If obvious fixes have not helped, the problem may lie elsewhere in the stack. 

 

 

What Counts as Good Drupal Performance? 

There isn’t one number that defines good Drupal performance, but Google does provide clear benchmarks: the three Core Web Vitals to measure different parts of the user experience: 
  • Largest Contentful Paint (LCP): Fast-loading content helps visitors get to what they need sooner. This score is good at 2.5 seconds or less. This measures how quickly the main content becomes visible.  
  • Interaction to Next Paint (INP): A responsive page makes interactions feel faster and more seamless. A good is 200 milliseconds or less. This measures how quickly the page responds when a visitor clicks, taps or interacts with it. INP replaced the First Input Delay in March 2024 and is one of the metrics where sites most often fall short. 
  • Cumulative Layout Shift (CLS): A stable page is easier to read and interact with; CLS is considered good at 0.1 or less. This measures how much the page content unexpectedly moves while loading. 
Google scores these criteria using data from real visitors and measures them on a page-by-page basis. Consequently, results from testing tools such as Lighthouse don’t always reflect the live-site experience, which means a fast homepage doesn’t necessarily indicate how the rest of the site performs.

 

  

The Three Layers Where Drupal Performance Problems Live 

Most Drupal performance problems come from one of three places: the application, front end or infrastructure. Identifying the right layer helps you focus on the fix that will solve the problem rather than spending on changes that will not solve the problem. 

 

The Application Layer 
This is where Drupal’s configuration and code process requests and generate pages before they reach the visitor’s browser. 
  • What lives here: Caching, modules, database queries, views and custom code. 
  • Symptoms: An issue in this layer can make the server slow to respond before anything appears on screen. This can happen when Drupal spends too long processing code or database queries. Logged-in users may notice it more because their requests are less likely to be served from cache. On medium and large sites, the database can become a significant scaling constraint. Drupal’s Core team notes that databases are harder to scale than adding additional web servers. 
  • Who fixes it: A Drupal developer can address these issues through configuration changes to rewriting codes. More complex issues may require Drupal development services. 

 

The Front-End Layer 
This is everything the browser has to download and render to make the page usable. 
  • What lives here: Image sizes and formats, CSS and JavaScript file weight, theme markup and render-blocking scripts. 
  • Symptoms: A problem in this layer can leave the server responding quickly while the page takes longer to become usable. Therefore, performance tools such as Lighthouse may report poor user experience metrics even when server response times look healthy. 
  • Who fixes it: A front-end developer, usually working on the theme, addresses these issues. For many Drupal sites, this layer contains some of the fastest and most cost-effective opportunities to improve performance. 

 

The Infrastructure Layer 
These are the hardware and services that keep the Drupal site running. 
  • What lives here: Hosting resources, PHP version, object caching (Redis or Memcache), CDN and the database server itself. 
  • Symptoms: The site may perform well under normal traffic but slow down under load. If pages are slow regardless of their complexity, the infrastructure may be the constraint. 
  • Who fixes it: DevOps and infrastructure teams investigate and address these issues, while procurement may be involved when additional hosting capacity or services are needed. 
The layer where a performance problem sits determines the fix. For example, more hosting capacity can help when infrastructure is the constraint, but it will not resolve inefficient Drupal code or a heavy front end. If the underlying problem is inefficient code, adding server capacity does not make that code more efficient. Diagnose the source of the problem first, then invest in the change that addresses it. 

 

 

Is My Drupal Performance Problem the Application, Frontend or Hosting? 

When a Drupal site is slow, the first question is which layer is responsible: in Drupal itself, in the browser or in the hosting environment. A few simple comparisons can help narrow it down before you change anything. 
  • Anonymous vs. logged-in: Open the same page while logged out and then logged in. If it is fast when logged out but slow when logged in, the application layer is worth investigating, particularly what Drupal’s cache is covering. The front end is unlikely to be the cause because both visits load the same images and scripts. 
  • Server response vs. render time: If the server takes a long time to start sending the page, the application or infrastructure may be responsible. If the server responds quickly but the page takes longer to appear or become usable, the browser may be doing too much work on the front end.  
  • Normal traffic vs. peak traffic: If the site works well with a small number of visitors but slows down as traffic increases, check whether the infrastructure has enough capacity. Higher demand can also expose application or database problems that do not appear under normal traffic. 
  • Before vs. after a change: If the site became slower after a particular change, start by checking what changed around that time. A new module, configuration change, code deployment or infrastructure change can give you a much narrower place to investigate than trying to optimize the entire site. 
Once you know which layer is responsible for the performance issue, you can involve the right team to investigate, scope the work and estimate the cost. 

 

 

Drupal Performance Tips That Deliver the Biggest Gains 

Not every performance improvement delivers the same return on effort. The tips below are ordered by likely payoff relative to the work involved, not by technical complexity, so you can focus on the changes most likely to make a measurable difference first. 
  • Configure caching properly, instead of just switching it on: Drupal has several caching mechanisms for different purposes. Internal Page cache serves requests from anonymous visitors, while dynamic page cache improves performance for a wider range of users. BigPipe helps pages appear faster by loading shared content first, then in personalized content later. This is why teams often enable caching and still see little improvement in their own experience of the site. 
  • Fix images before anything else on the front end: Large images can slow pages, especially when browsers download files larger than the device needs. Steps like using responsive image styles, switching to modern formats such as WebP and enabling lazy loading can reduce the amount of image data the browser needs to load. For many Drupal sites, this is both the cheapest performance improvement and the largest single front-end gain, yet it remains one of the most frequently overlooked optimizations. 
  • Audit modules against what they cost: Review each enabled module for the functionality it provides and the processing overhead it adds to page requests. If a module is no longer needed, remove it rather than leaving it enabled. Disabling a module isn’t the same as uninstalling it. So, switching it off does not necessarily remove everything associated with it. 
  • Keep the database from becoming the constraint: Old content revisions, log tables and inefficient queries can gradually increase the database’s workload. As the site grows, regular cleanup and query optimization can help prevent the database from becoming a bottleneck. Drupal support and maintenance service can help keep this work from building up again. 
  • Add object caching and a CDN when traffic justifies it: Object caching reduces repeated database work, especially for logged-in traffic, while a CDN speeds up delivery for distant visitors and absorbs traffic spikes. These are valuable scaling tools, but many teams invest in them before addressing less expensive application or front-end issues. 
  • Stay current on Drupal and PHP: Keep Drupal and PHP updated for the site to benefit from ongoing performance, security and support improvements. Drupal 11.3 is a good example of how platform updates improve efficiency. These improvements build on each other over time, but they do not replace fixing inefficient configuration or custom code on the site. 

 

 

When Does a Drupal Website Need a Professional Performance Audit? 

Most performance problems are configuration issues that a capable in-house team can handle. An audit is worth the cost when the problem is structural, keeps returning or cannot be clearly traced. 

 

Signs the problem has moved past configuration: 
  • It got slow after a migration or replatform. 
  • It’s fine day to day and falls over under load. 
  • It was fixed once and came back. 
  • Custom modules are involved that nobody on the current team wrote. 
  • Nobody internally can say which layer the problem is in. 

 

What to look for in a partner: 
  • They check all three layers before recommending anything: A partner who arrives with the answer already decided is selling, not diagnosing. 
  • They measure performance before making changes: A baseline gives you a clear point of comparison to show whether the changes actually improved the site. 
  • They rank findings by effort and cost: They show which issues matter most, what they will take to fix and what should be addressed first. A list of problems is useful, but a prioritized list tells you what to do next. 

 

At Material, these are the principles we use when approaching performance problems: diagnose before recommending and start with measurement to understand where the problem lies. Our Drupal consulting services follow that approach. If you are comparing agencies, our guide on how to choose a Drupal partner covers what to look for. 

 

 

Improve Drupal Performance With Material 

The first step in improving Drupal performance is finding which layer the problem sits in. Diagnosing the issue first helps ensure that investment goes toward the problem that actually needs fixing.  
With 200+ enterprise Drupal projects and 1,000+ Drupal Core contributions, Material brings deep Drupal expertise to both diagnosis and execution. Our work with Save the Children, for example, included addressing performance issues as part of its Drupal engagement. We focus on the underlying problem and the right solution without a hosting product influencing the recommendation. Get in touch with our experts to diagnose your Drupal performance challenges. 

 

 

FAQ: How to Improve Drupal Performance 

Why is my Drupal website still slow after enabling caching?

Your Drupal website can still be slow after caching because internal page cache mainly serves anonymous visitors, so logged-in users may see little improvement. Dynamic page cache can reuse parts of pages, while BigPipe loads shared content first, so personalized parts appear faster. Both need separate configurations. Caching reduces the cost of generating a page repeatedly, but it doesn’t fix the underlying performance issue and the first request after a cache clear still bears the full cost. 

Why is Drupal slower for logged-in users than anonymous visitors? 

Drupal is often slower for logged-in users because personalized pages cannot use shared cache, so each request requires more work. BigPipe loads shared parts first to make personalized pages feel faster. This difference is most noticeable for editorial teams, whose day-to-day experience of the site, as logged in users, is very different from that of anonymous visitors. 

Can Drupal improve performance without a full rebuild?

Yes. Drupal sites and apps can usually be made faster through configuration, image and asset optimization and targeted infrastructure changes. A rebuild is mainly needed when the content structure or custom code is the bottleneck. Core updates can also improve performance without changing site content or structure. 

Why did my Drupal site become slower after an update, migration or new module? 

If your Drupal site slows after a specific change, start by finding what has changed. Changes to modules, code or configuration can alter the work Drupal performs on each request, while migrations can change hosting, PHP versions and caching at the same time as the code. Establishing a clear timeline helps narrow down the cause before you start optimizing more broadly.  

Can Drupal handle sudden traffic spikes without slowing down or crashing?

Yes, Drupal can handle traffic spikes with edge caching and enough server capacity. Anonymous traffic is easier to serve from cache, while logged-in traffic puts more demand on the application. We recommend doing a load test before a known launch, campaign or press event to identify potential performance issues before they affect users. 

How should Drupal performance be monitored after the site is optimized? 

Performance improvements may not last forever, so monitor Drupal performance regularly as the site changes. Track Core Web Vitals from real visitor data, server response time and page weight as content and modules grow. This helps catch performance drift before it becomes serious.