Monitoring a Web application with ServiceNav

In today's application usage, web applications can represent an important part of an information system.

Whether it is transversal services (extranet, messaging, etc.) or business services (SaaS CRM, e-commerce platform, dematerialized billing, etc.), their availability is often critical for users and your business.

It is therefore more than relevant to supervise, whether they are SaaS or internally hosted applications.

Choose the right level

Two levels of supervision are possible:

• Simple Hardware/system infrastructure, access to the URL.
• Complete : above scope + application scenarios, control of their result and their execution time.

The choice of a complete or simple supervision will be influenced, for each web application, by :

• The criticality of the application for the proper functioning of your business
  • What is the impact of a slowdown? a production shutdown? are the users only internal or is it a customer site? Is it a test environment? 
• The associated ticketing history
  • Is this web application the source of many support tickets? If so, what is the majority of them?
    • login page out of order -> a simple supervision is enough, with a URL test.
    • slowness in creating a record or search -> supervision by scenario necessary.
• Available resources
  • Creating, and especially maintaining, test cases takes time, as it is necessary to ensure that they remain compatible as web applications are upgraded.

The steps

As always, we start by breaking down the service provided and assessing its criticality.

For example:

• Services : CRM website
• Hosting on site
• Architecture 
  • 1 Windows VM: Web Server
  • 1 Windows VM: Sql Server database
  • Access URL: https://moncrm/login
• Criticality
  • Used by the sales team and helpdesk + extranet part for end customers
  • => Review
• Ticket history
  • Rather stable and fast tool
  • Sometimes the page is inaccessible when the application server disk is full

We will therefore choose here a supervision simplebecause the URL test is considered sufficient and reliable.

Simple supervision

Add (manually or by census), using the equipment/service templates appropriate to the role:

• The application server
  • For example, the service model HTTP_HTTPSwhich allows to test the accessibility of a URL, and the validity of an associated certificate
  • It is important to also monitor the IIS service via the MS-WIN-ServicesList-Started
  • Special attention will be paid to the disk, as it generates support and downtime, for example with MS-Win-DiskUsageA warning threshold is set according to the growth rate of the disc.
• The database server
  • The SQL Server database can be supervised with one of the many models available in the catalogue.

You can also use other models adapted to Windows servers, in order to monitor the CPU, the RAM, the disk I/O for example. These models are associated by default with the 'server-model-windows' equipment model.

Complete supervision

We could just as easily have chosen to implement one or more application scenarios. The subject is detailed in a dedicated blog post.

And for a SaaS application?

In the case of a SaaS application, the control points related to the server will not be relevant, the device can be configured as 'Not pingable', being associated with the controls of the different URLs accessing the service.

The general philosophy then remains the same.

IT Weather Services

Finally, we will ensure a synthetic and efficient follow-up of our newly supervised application, through the construction of a service weather.