Operations | Monitoring | ITSM | DevOps | Cloud

The past month we dedicated time and resources into optimising the speed and experience of our public website, our dashboard and our behind-the-scenes uptime checks that we perform. Overall, our website and dashboard feels about 2x to 3x faster. The biggest gains are for our users that have > 100 sites on their dashboard, they'll get a noticeably faster loading time. For those biggest users, the dashboard is quite litterally 10x faster.

As part of our big deploy that added ping and TCP monitoring, we’ve also shipped a small, but often requested feature: you can now choose the frequency of the check we run. By default, we check your website for uptime every minute. The Lighthouse check runs daily. Using our new feature, you can now, for instance, choose that the uptime check should run every 2 minutes, and the Lighthouse check every 5 days. You can choose the frequency at the settings of the check.

A couple months ago, we sent out a survey to all our users asking what they like about Oh Dear, how they use it, and how we could improve our service. One of the things that was asked a lot was ping and TCP port monitoring. The past few months we worked hard to add this kind of monitoring to our service. And while building it, we touched upon other parts of our service and improved lots of little things. And I'm proud to share that we now have shipped it all! Let's go through it!

Today, next to Ping and TCP monitoring, we've also launched a new PHP SDK package, which has been rebuilt from scratch using the wonderful Saloon library. Using our new SDK, you can easily use the entire Oh Dear API. In this blog post, I'd like to show you how you can use the new SDK and how it works under the hood.

Whenever you visit a website or click on a link, there’s a whole conversation happening behind the scenes between your browser and the web server. That conversation includes something called HTTP status codes and knowing what they mean can help you make a diagnosis, so to speak. Usually, everything goes smoothly (like a 200 OK), but sometimes things break (looking at you, 404 and 500).

One of the core features of Oh Dear is that we can notify you whenever we detect problems with one of your sites. Our notification system is quite powerful. We support many different channels (like email, Slack, Telegram, ... and a whole bunch more), and have fine-grained control over which events should trigger a notification. Today, we've added notification logs.

Whether you're keeping a personal blog or manage a corporate site or online storefront, website downtime can cost money and can damage your reputation. Let alone when you're maintaining a bunch of different client sites. And while downtime can't always be prevented, it's really easy to at least keep track of things, and diagnose potential issues from there. So, let’s start with the easy part.

SEO is in a weird place right now. About one in five LinkedIn posts in my feed currently claims that SEO is dead, or has been assimilated by LLMs. Do not be remiss, dearest reader, because even an LLM still uses search engines like Google and Bing for web crawling. In other words, SEO still matters, a lot. Additionally, it's never a bad idea to keep tabs your website's SEO performance.

A critical security vulnerability was recently discovered in Livewire v3 that allows remote code execution, as Stephen Rees-Carter reported on Securing Laravel. While patches are released quickly, many applications remain vulnerable because developers simply don't know about the issue yet. Oh Dear's Application Health monitoring solves this by continuously checking your production environment for security vulnerabilities and immediately notifying you when issues are detected.

When you're managing a handful of client sites, setting things up manually is fine. Though if you're managing dozens of them, you're going to think twice about your approach. For agencies, development teams and platforms who are responsible for loads of websites, having to repeat the same configuration over and over is not only inefficient but also more prone to errors. That’s where this blog post comes in handy.