Web Performance Monitoring Metrics

We previously discussed 5 different types of web monitoring. Here is the list of important metrics that each of these services can provide for holistic web performance management.

Visitors Tracking

• Pages views
• Unique visitors
• New visitors
• Returned visitors
• Visitors distribution by countries and cities
• Browser statistics
• Max traffic
• OS statistics
• Referrers
• Entry pages
• Exit pages
• Time spent
• Conversion rate
• Most popular pages
• Common paths
• Search keywords

External Monitoring

• Response time from different external locations
• Uptime daily/monthly
• Average response time all location per URL
• MTTF (mean time to failure)
• MTTR (mean time to repair)
• Aggregated Uptime
• Incidents after working hours
• Average response time all web services
• Planned downtime
• Failures and notification

Internal Monitoring

• Server resource utilization (CPU, Memory, Network, HDD, Virtual Memory, Threads, Process)
• Total free disk space
• Number of alerts on exceeding system capacity thresholds
• Process resource utilization
• Response time local
• Process uptime
• Server uptime
• Uptime local

Online reputation Monitoring

• backlinks per SE
• indexes per SE
• position per keyword per SE
• social bookmarks
• references in blogs
• references in news
• page ranks

Transaction (application) Monitoring

• Failed transactions
• Coverage
• Total execution time
• Execution time per each step
• SLA figures (uptime)
• Page load breakdown
• Average execution per step

How Web Developers Can Keep Customers Happy

Website designers and web application developers provide support for the sites they developed. Most of the companies are small and don't have large marketing budgets, so they relies on word of mouth generated by happy customers. Developers typically either host or have VPN access to customer sites and when customer experience problem with the site they call web developer and get problem fixed. And customer satisfaction very much depends on their website uptime and service level provided by the web developer.

There is way for developers to improve customer satisfaction. Using Monitis internal monitoring, web developers can setup monitoring agent and perform customer server monitoring right from their location. By checking all the nodes between end user and server they got complete and exact picture and early notification about current or upcoming problem.

Combined with external monitoring, internal monitoring may provide more proactive way to handle problems. Even if it is intranet application, developers can use VPN connection for monitoring the application. When customer calls developer he will be please to hear “yes, we know the problem and already working on it.” Or developers may call their client and say "we notice an issue on your server and will fix it now." Moreover, by monitoring nodes, developers may quickly identify problem and fix it, thus increasing their productivity and reducing mean time-to -repair (MTTR).

5 Important Website Metrics Business Should Monitor

1. HTTP Uptime. HTTP uptime monitoring is important because each time website is down it a) reduces website ranking by search engines b) reduces website traffic and income c) waste PPC budget

2. External HTTP response time. Response time is a measure for user experience. When website is slow it significantly reduces user satisfaction.

3. Web Traffic Statistics. It is important by itself to analyze website visitors demographics, used browsers, referrers, etc. But it can provide more insights if combined with other metrics. For example web traffic measurement together with HTTP checks may help to find load problems when site performs slow in case of increased traffic, or together with search engine monitoring it will help to understand who is referring to your site and how much the traffic depends on the position on Google and other search engines per keywords which drive traffic towards your site.

4. Search Engine Position. Internet marketers need to watch daily website position in major search engines for sometimes hundreds of keywords which relates to website content. It helps to fine-tune search engine optimization campaigns and do competitive analysis.

5. Social Media references (Buzz Monitoring). Social Media and book-marking sites are getting more importance as a way to bring more traffic and conduct online PR campaigns for building Internet brand and reputation.

Ideally you need to have all these metrics real-time at the same dashboard in order to have full picture of your website performance. Monitis provides on-demand http and web traffic monitoring; Semonics service from the same vendor provides search engine and social media monitoring, MyBuzzMonitor could also help to stay informed and also show your reputation to your website visitors via live widget you can place into your blog or web page.

Merging of Web Analytics and External Monitoring

Good blog post A view of the web analytics market. In the author words:

My take is also that monitoring will eventually be much closer to web analytics ... and end-to-end monitoring of the user experience will play a bigger role. It's obvious that a degraded web performance or poorly performing enterprise systems have a direct impact on conversion and outcomes on the frontline. Unless I'm mistaken, this is not measured by any of the ASP-based solutions.

That is actually what Monitis website performance monitoring service is providing. Having traffic and response time charts helps to notice a degraded web performance correlation with web load.

5 Reasons for Merging Internal and Exteral Monitoring

Most on-demand website monitoring service providers only provide external website monitoring. Althoug being an important performance control service, external monitoring alone cannot identify the root cause of failures. Merging internal and external monitoring data provides the following benefits:

1. Quickly indentify and fix the cause of failure or slow response, thus reducing site downtime.
2. Fix the issue before it affects end-user experience
3. Understand internal and external user experiences
4. Check network latency and connection failure between you and your business partner and dependant web services.
5. Find causes of past failures by analyzing historical data and plan appropriate infrastructure investments.