Monitor server load time

The speed of a website is one of the most important quality criteria. Should an online store, a company website or an editorially maintained site is slow or perhaps only appears so, this can quickly scare away potential customers or users.

What is the server loading time?

The speed of a website is made up of two components. The browser load time and the server load time. So the time that a browser, such as Chrome or Firefox, to render the content and make it available to the user, and the time it takes for the server to render the to generate the HTML for the browser. The server load time is also known as "time to first byte" (TTFB).

Why is server load time important?

There are a few reasons why server load time matters. We find the following three to be the most relevant.

First and foremost, let's put ourselves in our customer's shoes. He is used to Google, Amazon and Facebook speeds. If we deviate from this in a negative way, we will lose a few points in terms of trust. In e-commerce in particular, we have to put up with comparisons time and again. comparison. Perhaps that's not always entirely fair, since the world's largest online retailer has thousands of developers under contract. Nevertheless, they do they do the same as the online retailer around the corner: they sell products to the end customer.

Speed is also becoming increasingly important for search engines. On the one hand, because they want to recommend user-friendly websites rather than those that cannot be used properly, but on the other hand - and this is perhaps the more important of the two points - they have to index a website so that they can index a web page so that they know the content. Since most queries by search engines do not go through the browser the crawl speed depends solely on the server load time. If the server is slow, Google and Co. will pass by less often or it will take longer until all pages are found in the search index.

From a technical point of view, there is another, very relevant, point. Scalability. Time that happens around browser, happens in the individual browser of the user. That scales quite well, because we have one browser per user and one computer behind it. A thousand users, a thousand computers. With the server, it looks quite different. A thousand users, one server. Sounds a bit more unfair. The shorter a piece of software takes to compute the HTML, the more requests it can answer. And the cheaper it is to run a high-traffic website. The greater the server load time of a website is, the higher the chance that it will eventually break down under heavy load. In summary, these three points are important for the success of a website:

• Performance as a form of usability
• Search engine friendliness through better indexability
• Lower server costs

.

Who influences server load time?

The biggest part of the load time that an application needs on the server depends on the software that is used. Unfortunately software used, but it is not always in one's hands. For example, a store based on Shopware cannot be faster than a new installation of Shopware. The good thing is, most standard applications are already fast, otherwise they would probably not be used by so many users.

Those who use software in its pure form do not have to worry about speed in most cases. But hardly anyone does. We know almost no store that manages without a single plugin. We all want to stand out a little bit and that is not possible without individualization.

In addition, the performance of the server plays a role. Fast server, with many cores means a higher chance to deliver the page to deliver the page quickly, if several users are on it at the same time.

So, in summary, three points are most relevant:

• System used
• Plugins used
• Server performance

How do we measure server load time?

koality.io measures the server load time (time to fist byte) every hour. It is important to know that our monitoring tool, the web pages through a real browser (Chrome). The server therefore always delivers the page as if a real customer were browsing it. For the measurement, we use the so-called Navigation Timing API, which is is now a standard feature in all major browsers.

.

What comes closest to the time to first byte here (DNS and co. are comparatively fast) is the responseStart value, which we also use for this also use for this. After that one gets all measured values from the diagram displayed.

Important for continuous monitoring of such an important core property are the historical values. These help to recognize whether a website is still fast, but it has become slower. This can happen, for example, due to the installation of a new plugin or simply by the increased load on the server.

But not only the historical values have a relevance, but also a wide collection of the values. As described at the beginning, the performance depends of the website, depends on the used framework, CMS, blog or store system. But not every page in such a system has the same speed. A category page, for example, is much more complex to calculate than a simple imprint page. Much more database query and logic is needed here. That's why it's important - and this is how koality.io solves it - to always have a reference page for each page type of an online offer with in the monitoring (equivalence class principle).

How can I optimize them?

Optimizing server load times is one of the standard tasks of any web developer. Here are some ideas what you can spontaneously can do to reach the goal quickly:

• Faster Server - Slow web servers can negatively impact website performance. Here you can for relatively little money faster

systems for relatively little money.

• Look at database queries - Much of the time an application takes to run will happen in the database. Here it is worthwhile to

For example, it's worth familiarizing yourself with MySQL's Slow Query Log.

• Turn off slow plugins - Often it is plugins that have a great feature set, but are not programmed to perform well,

that slow down entire systems. Here you should check if there are alternatives or if you can do without the plugin.

• Caching** - Caching, i.e. the caching of results to ensure fast access, should always be active.

Here, most systems such as Wordpress or Shopware already bring native implementations, but it is worthwhile to look at plugins on the topic of other providers. Often these optimize the speed even more.

• System updates** - Shop systems, CMS, blogs - all of these are continuously evolving. Not only in the range of functions, but also in

speed. So it is advisable to update all used systems to their current version to benefit from the innovations.

• PHP-Update In our article, why PHP is the best language for agencies, we already described that

we already described that PHP is the programming language on which many of the most commonly used CMS and store systems are based. In the last In recent years, the language has developed very strongly, especially in the area of speed. Here an update is worthwhile in any case.

Conclusion

The server load time is one of the most important properties of a web application. It should be monitored continuously, across all page types. across all page types. Historical values help to detect changes over longer runtimes. In the best case, optimization can be done by simple changes can be made.