Downtime is a big issue for websites. For SaaS companies that are happy that their services are available, even a few minutes of downtime can mean a direct loss of revenue. Here’s how to track if your site is not working.
How to track downtime
The problem with downtime is that it is difficult to track yourself. After all, you̵7;re probably not updating your site every ten seconds and waiting for it to go offline. The best solution is to have a downtime monitoring service that tracks your website for you.
These services periodically check your website and trigger warnings if something does not work properly. Because problems can be located in different parts of your site, most downtime tracking services will check multiple pages and usually smooth out the free and paid plans based on how many checks they will make.
If you have some form of static caching in front of your server (like Amazon CloudFront), your services may go down with the website as normal. In this case, you want to check a page that is not cached regularly and must retrieve information from a database. Some of the tools listed below can even perform content checks, which will verify that your web page is intact and looks as it should.
How do downtime screens notify you?
Page downtime is a major problem, so it’s important that you get a message quickly so you can fix the problem. Most services can send emails, text messages and Slack messages, which may be enough for many people.
But if you do not actively look at your messages, you may miss them, which can lead to your site being down longer than it should. If you need to be notified of major problems as soon as possible, we recommend the Uptime Robots Pro plan, which can be configured to repeatedly call you if your site is down for more than a certain amount of time.
They will call you to verify your number when you configure the voice call service, so that you can whitelist the number so that it still rings even when Do Not Disturb is activated.
Voice calling, however, costs money, so it is not unlimited; you have to buy SMS credits separately. They include some in all Pro plans, but you only get them once and they do not renew.
Options for monitoring downtime
Uptime Robot is a value monitoring service. The free version allows 50 monitors at 5 minute intervals, but requires $ 5 per month Pro version for text messaging. The Pro version also includes voice calls, which can be repeated periodically to ensure that you receive the message.
Uptime.com sends SMS for $ 8 per month. It can perform content matching, API calls, as well as a service called “transaction monitoring”, which can, for example, try a login flow from a browser to verify that users can log in.
During heavy traffic, your site may remain functional and visible but have significantly slower loading times, which is still a major problem. Pingdom is a performance monitor and analytics package that can notify you if your site is starting to slow down and can often tell you what is causing the slowdown in the first place.
CULA is a free tool that can test if your website looks like it should. It can monitor 50 URLs with 5 content controls per URL and can send mobile push notifications with Pushover (but no SMS). Their paid versions allow more URLs, but the free version is very generous.
AWS CloudWatch is a complete monitoring suite for AWS products. It does much more than monitor your site’s uptime, such as collecting logs and monitoring your overall infrastructure, but can be configured to track the performance of your services.
Open source options (host for yourself)
These tools are open source, which means that there is a package available that you can install and run on your own server. We recommend that you do not install monitoring software on the same server, because if there are problems with the server itself, there will probably be problems notifying you. The best solution would be to rent a small server from a completely different cloud provider, to significantly mitigate this risk.
Zabbix is a complete monitoring suite that not only tracks the uptime of your website but also tracks the uptime of your applications and backend services.
Icinga can monitor many services and generate regular health reports on your servers. It can notify you via email or SMS if any critical issues should arise.
Prometheus is a surveillance suite, built by SoundCloud before it goes open source, which works especially well with orchestration engines like Kubernetes. It provides easy monitoring of your Kubernetes services and can notify you of unexpected behavior.