High Load on Websites: Why Is It of Importance Not Only for Programmers?
Subscribe to our blog to receive recent updates
Whether you’ve pulled off a great ad campaign or published a super-topical article which made the first few pages of search engine results, one of the measures of your success is an increase in relevant traffic to your website. The notion of increasing traffic to their site makes product managers and service marketers happy: their thoughts immediately turn to increased revenues from advertising, growth in sales (if this is an online store or media with a paywall), and more social mentions and repeat visitors. Could there be any downsides? Well, yes, there could, if your service isn’t ready for the increased load.
Every user who navigates to your resource and clicks on something, or even just opens a page, creates a load on computing power. If this load on the server increases, and you haven’t made provisions for flexible scaling of your server capacity as part of your project, your equipment ends up pushed to its limits. The result is that your applications’ performance suffers: in layman’s terms, your website or app is slow. You can’t define “high load” with a single number: this is a problem that could affect both a little WordPress blog and a giant like Facebook. The critical number of users will be different for different projects.
What are the Signs of Excessive Load?
To make the phrase “high load” less abstract, let’s take a look at the issue through the eyes of a user: pages load slowly, you get 502 and 504 errors and database errors, pages don’t open, submissions don’t get saved. Someone looking at the site analytics at that moment would see the page views per visitor falling, the bounce rate increasing, and the site’s search engine ranking decreasing. A more specialised metric which engineers might notice increasing when a site is slow is TTFB, or time to first byte. This is the time that elapses from the moment a browser starts to send a page’s request to the moment it receives the first byte of information back from the server. Poor TTFB is a pretty sure sign of performance issues.
What Can be Done to Avoid Problems?
We’re talking about the kind of load generated by service users, and most of this is predictable. Given the consequences and losses that could result from the website not working as it should, the team responsible for user interaction should always assess whether any actions they’re about to take will generate a spike in traffic. If you’re preparing an ad or SMM campaign, having a sale, or sending out a newsletter which will attract a lot of traffic to the site, tell your engineers what your plans are. Together, you’ll be able to assess whether the increased load you’re expecting will be critical for your system.
There are a few ways you can increase the load your site can cope with. You can scale up your system by increasing its computing power, i.e. buying new servers or upgrading your web hosting plan. Another approach would be to speed up your system by adjusting its cache settings and optimising the code it runs (you can read about the many and varied things caching can do here). The solutions you employ can vary depending on the volume of traffic you’re expecting, the type of load (one-off or gradually increasing), and your site’s most frequently-used functions. Increasing your site’s resilience isn’t just a problem for engineers: there are also economic considerations. As your project grows, the cost of scaling increases. To reduce this cost, you should hire experienced engineers and carry out a thorough analysis of your system.
Adapting your system to increasing loads is a natural stage in a digital project’s evolution. This article hasn’t touched on the kind of increased load that can occur for reasons that wouldn’t be on a non-techie’s radar: security issues, web crawlers, suboptimal database setup, etc. If the engineers attached to your project are doing their job, these kinds of issues will be resolved without affecting other people involved in the workflow. At the same time, though, to avoid the problems this article has discussed, businesses should forecast their growth so the necessary tech can be put in place in good time.
