At the start of the year we began the ambitious task of revamping the tools we offer our developers and customers. We’re all designers or programmers at heart, so we started to rethink how to make server monitoring more accurate and more usable.
Server monitoring has been the achilles heel of many hosting and infrastructure companies. Sure, many are able to provide a basic monitoring window that alerts you to any server issues, but the problem of multiple systems and multiple logins remained the biggest unneeded complication.
What we're doing
Our goal for Central is simple, we want to provide great server monitoring while keeping the system usable for the techie and non-techie alike. From a single login and single dashboard we will be able to provide a converged view of your system, application and business metrics. But what does this really mean? From Central, we’ll show you how your application, server and business are performing. For anyone who’s tried to combine all of this before, it’s no easy task!
As most developers know, most of these tools already exist. You can monitor server performance with the Plesk Health Monitor, Munin, PRTG, Nagios and many other tools. We don’t want to replace these tools, instead we’ll be collecting the data into a central database and single. The amazing monitoring these tools provide will remain, but combining it with data on how your application and/or website is performing is the key.
Let’s jump into Central and look at some examples. After logging in with your Conetix Control Panel details, you’ll be presented with the Central Dashboard. Here you’ll be able to identify any container issues immediately while getting a greater insight into the general health of all your servers.
But what if you see an alert about high CPU usage? What impact is it having on your website? Are you seeing more errors? Are you seeing longer page load times? Is it having no measurable effect? Playing detective when sites are down is far from fun. With Central you’ll be able to quickly drill into the server having issues and discover any unwanted performance patterns.
Conversely, what if you start to see longer page load times, what is your server doing at that time? Did you see a spike in connections? Did you see a spike in processes running? What other events correlated with your longer load times?
Taking this a step further, we want to give you a broad overview of how all your services with Conetix are performing, which is especially handy when you have dozens of Virtual Private Servers (VPS’s) like many of our larger clients do.
The First Steps
We’re going to have the first steps ready for beta at the end of next week, and we’re already excited. We’ve focussed on getting the base wireframing of the system right, which has focussed on a few key areas. Our initial release of Central will be aimed towards the developers, with the intent to build a fuller product for business owners soon after.
Secondly, we’ve put a core focus on scalability. Our aim to build Central in a platform that will handle a lot of data (we really do mean a lot), yet remain very responsive.
Thirdly, we want complete automation. From an end user perspective, this is a near zero configuration system. Unlike other monitoring where it needs to be configured, we have built a platform which draws data from our billing system and then uses it to automatically build a list of your servers to monitor. All you have to do is login!
At this stage, we’ve implemented the Linux VPS monitoring only, but unlike other systems we’ve added some deep monitoring to provide useful data. For example, we have CPU latency measured for each VPS, which provides a far more useful value over CPU utilisation. The processing latency gives you a real indication of the system performance as it’s a measurable impact on what’s going on.
What's Next
We have some big aspirations for this project over 2014. Our aim is to make the system so immensely useful and powerful that it’s always open. Over the next few weeks we’ll finalise our current roadmap and push it out for feedback. We genuinely want you involved, so please feel free to provide suggestions and feedback.
Until then, he’s a few key areas we’ll be focussing on:
- Website Monitoring. This is a critical part of monitoring server performance and while it can be done with other tools, we’re making it central and automated. We want the list of websites monitored to be automatically generated and tracked between server moves.
- Business Metrics. This will focus on areas like Magento to gather transactional information and show a combined view with your server and website performance.
- Custom Wallboard. We’ll be allowing you to customise what data you want displayed on a big screen in your office. You’ll have your own mini network operations centre!
- Application Metrics. Rather than just server performance data, we want to provide monitoring and statistics on systems such as Apache and MySQL. Again, this will be a zero configuration option, it will simply occur via automation.
- Application Logs. Ever had to dig through an apache log to figure out what occurred when a server had a major load spike or the website response times skyrocketed? We’re aiming to have this data automatically correlated so that all you have to do is click on the spike.
