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

5 Monitors for Ultimate Business Performance: Part 5 Transactions Monitoring

5. Transactions (a.k.a. Application, Web Workflow, Functional) monitoring – monitors not only single web page but sequence of web screens and user actions. It allows checking functionally of web application logic and measure the performance of web transactions. A test script should be developed in order to perform such checks, either by automated recording of user actions or by writing the code manually. Monitis recently beta launched a transaction monitoring feature which allows to record, edit by hand, and debug test scripts. After recoding the script can be deployed on the server and then will be running on regular intervals. Users can see total execution time of script, but also can drill down to execution time per each step. In case of failure Monitis will send an alert. It also records screens which did not pass via checkpoints, thus allowing understanding the problem of failure later on. Finally it provides breakdown of load time per each component of the tested web pages.

Who should use it: business owners, webmasters, IT operations, QA and web developers.

Usage examples: a tool for regression tests of live production or pre-live development servers, functional monitoring to make sure main application flows are functioning, performance management and tuning.

5 Monitors for Ultimate Business Performance: Part 4 Search Position Monitoring

4. Search Position (a.k.a. Search Visibility, Online Reputation, Search Ranking, Site Popularity) monitoring – monitors website position on search engine result pages (SERP). Search engine optimization work requires monitoring position of website URL per many keywords, sometimes several hundreds, and per several search engines like Google, Yahoo, MSN, Ask and others including country specific ones like Baidu in Chine and Yandex in Russia. Doing manual checks is time consuming and doesn’t show historical progress. There are several expensive software available to automate this task, but there is a new on-demand search engine monitoring service which can do the work on monthly subscription bases. It also monitors social media and bookmarking sites.

Who should use it: business owners, search engine marketing and optimization consultants, webmasters.

Usage examples: progress monitor and reporting tool for search engine optimization (SEO) campaigns, competitive research, online reputation and buzz monitoring, brand management.

5 Monitors for Ultimate Business Performance: Part 3 Web Traffic Monitoring

3. Web Traffic Monitoring (a.k.a. Visitors Tracking, Hit Counters, Web Analytics, Web Statistics) – monitors website page views and visitors. It is possible to monitor web traffic via 3 different techniques: a) by including visible or invisible html code into web pages which sends requests to an external counter b) by using web log analyzers like open source Webalizer or c) by using network sniffers. The last is quite rare and only used by large Internet portals. Now html code options becomes more popular especially when users don’t have access to the server for example in case of blog service providers.

Who should use it: business owners, IT, web marketers, webmasters, web developers.

Usage examples: Demographics of the visitors may provide better understanding from where users visit the website. Unique visitors trend also shows how well is site promoted. Referrers and keywords analysis are useful for search engine optimization and search engine marketing. QA and Web developers may perform load analysis and also find most popular browsers and OS so they can check and optimize the site for these platforms.

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.

7 Website Checks to Ensure Reliable Service

Most people only understand http ping checks by Web site monitoring. Monitis premium website monitoring provide options for deeper and comprehensive checks. In case of running business critical website, it is important to apply these checks in order to reduce downtime, increase web traffic and search engine ranking. Here is a short checklist:

1. Add ping test to your Server
2. Add Http tests to Home Page.
3. Add HTTP checks to other critical pages especially with dynamic Content. Especially important to add URL checks for password protected pages. Advanced Monitis HTTP/HTTPS checks allow to check also password protected pages and form submit reply pages (e.g. using POST). So for example if it is web mail, rather than just check home page, site owner may create a demo user and check page validity after login.
4. Add ping tests to firewall, router and other nodes between your website and public network. That way you can quickly find the issues. In case if your router disconnected then there is no need to check website. You can find intermediate node IPs via for example traceroute. You can check your network devices using ping or telnet protocols. Some devices also will provide HTTP interface so you can check also via HTTP.
5. Add DNS checks for your domain. In same cases your website could be completely reachable although your lose visitors because of issues on the name server.
6. If you have access to your web server then download and install internal Monitis agent (Linux or Windows) and add CPU, Memory and HDD checks. You can also add checks for Web server, and Database process. In that case you can see your server bottlenecks and fix them even before your website goes down. You can also add http checks within your firewall in case of web services which are not visible from outside.
   
7. If you have Database you can add external Database checks. Just make sire you enables access from our monitoring IPs.