Many companies are dependent on a high-performing, available website to conduct business. Whether it's an online store, a landing page for customer acquisition or online support, web performance is critical to business success. Downtime means lost dollars, and long-term problems can put the business at serious risk. Some estimates have put the cost of downtime and outages into the hundreds of billions per year.
IT departments often try to protect against downtime by focusing on the web application. Monitoring web application's performance helps identify malfunctions and their cause on a code level, so that the DevOps team can solve the problem. But, monitoring application performance only protects against application errors and ignores external factors such as network traffic, hardware, connectivity issues or bandwidth usage, all of which can have an impact performance and availability of a website.
When website performance is poor, any individual component can be responsible. Worse, the search for the root cause can be time consuming and difficult. The best way for IT departments to approach this specific problem, therefore, is not to focus on which point solutions solve specific problems, but to engage in preventative maintenance of all systems. If the systems administrator constantly monitors all of the components involved in a website process, they can baseline normal patterns and set ranges that alert to anomalous behavior.
Collecting that type of data is extremely useful in anticipating issues and identifying them before they become problems. The main goal for IT in this instance is not to establish efficient backup and recovery processes, but instead, to prevent the types of issues that lead to failures and outages altogether. Additionally, over time administrators can optimize systems and process based on historical data, which only increases the resiliency of the website and enhances overall performance.
Administrators looking to monitor website health performance in a more holistic way need to find a solution that can comprehensively monitor all aspects of the IT environment. To monitor the website end-to-end, IT would have to take the following steps:
1. Website monitoring via ping
2. Monitoring page load times
3. Web server monitoring (Microsoft Internet Information Services IIS, Apache, nginx)
4. Transaction monitoring
5. Out-of-box monitoring of common devices and applications, such as servers, switches, routers, databases and firewalls
6. Support of standard protocols for monitoring data streams such as SNMP, NetFlow and Packet Sniffing
7. Monitor virtual applications
If an administrator can put in place a comprehensive monitoring strategy that can track every aspect of the website process, they will be able to identify issues before they become problems, decrease downtime, and protect a mission-critical business process.
Dirk Paessler is CEO and Founder of Paessler AG.