User story : Sanity check - easysoa/EasySOA GitHub Wiki

<<< Back to [[Final soa model design]]

• Introduction
• User story
• UI requirements
• Model requirements
• XXXs: Missing information
• (Draft)

• Role involved: SOA Architect, Administrator, Developer
• Goals
  • Get a map of the dependencies between services
  • Be able to know when the services I use but have no control over change (XXX1: What about the other way?)
  • Be able to ensure the quality of services by running tests
• Features
  • Cartography of the SOA: dependency relations with external services
  • Web services validation
    • Various types of validation
      • Availability
      • Unit tests (request/response)
      • WSDL validation
    • Notifications/reports when validation fails
    • Validation scheduling

As a SOA Administrator/Architect...

1. GIVEN that I have a complete model of my services (and the services I use) in EasySOA

• AND I need to gather exchanges to set up request/response validation tests
• WHEN I use the service registry to browse to the service I want to test
• AND I start the client scaffolder
• THEN a client is generated for the service

2. GIVEN that I have started a scaffolded client for my service

• WHEN I make one ore more requests
• AND I choose to store them
• THEN they will be kept for later use (for validation tests)

Monitoring case:

3. GIVEN that a probe/proxy has been set on my web service

• WHEN a requests is made to this service/through the proxy
• THEN the request is stored for later use (for validation tests)

As a SOA Administrator...

1. GIVEN that I want to set up some validation for my web services

• WHEN I open the validation tool
• THEN I am invited to configure a new validation configuration

2. GIVEN that I am on the validation configuration screen

• WHEN I choose some web services to target (e.g. an external "Weather" service)
• AND I choose the validation features to enable (e.g. availability check + request/response tests + WSDL consistency over time test)
• AND I configure them (request/response tests : choose the set of requests to use, WSDL consistency : choose the validation stricness)
• THEN A new validation configuration is created and made available

3. GIVEN that I want to test my newly created validation configuration

• WHEN I manually trigger the validation
• THEN the validation is run
• AND a validation report is produced

4. GIVEN that I want to schedule this validation configuration to run regularly

• WHEN I configure the schedule
• AND I configure the notifications to trigger (send what? to who? up to which frequency?)
• THEN the validation scheduling is set up

1. GIVEN that a scheduled validation is run on a certain webservice

• WHEN that service fails during validation
• AND a report is produced
• THEN a notification is sent to all users chosen during validation configuration

As a developer...

1. GIVEN that a service A failed during WSDL consistency validation

• AND I have been notified as the developer of the service B that requires it
• WHEN I fix and redeploy my service B and make sure everything works again
• AND reset the WSDL consistency test for service A
• THEN the validation for service A works again, and uses the new WSDL to test consistency

As an architect...

1. GIVEN that an external service failed during availability validation

• AND I have been notified
• WHEN I contact the provider of that service to solve this issue
• AND the service is back to normal
• AND I run a manual validation of that service
• THEN the service passes again the validation

• Services browsing
• Scaffolded service client
• Services validation
  • Service selection
  • Validators selection & configuration
  • Validation scheduling
  • Notifications configuration
  • Validation reports & mail notifications
• Validation dashboard (view the state of the services)

We need to be able to store:

• Web service exchanges (outside the document model?)
• Services (their names, URLs, the URLs of their EasySOA proxy(ies))
  • Service validation configurations & schedules

• XXX1: When my own web services change in production, do we need the capability to notify users of these services?

As an SOA architect or operator,

• I want to compare how my own unchanged SOA is linked with and behaves with an external, unknown SOA to be sanitized, so that I can detect changes in link and behaviour
  • (sanity check ("validation") dashboard ; in 0.4, see wiki doc)
• I want to be notified when a service definition (WSDL) changes on an external provider's endpoint, so I can ask its provider about it (in nuxeo or not) or patch my code (possibly in a proxy)

As a test maintainer,

• I want to be notified when a functional test involving a given external endpoint fails...

=> after auto sanity check, notify (send a link to the report saved in nuxeo to) of errors (to be found in the report) interested people (model : link to owners of endpoint or tests...), jenkins-like

As an SOA architect or operator,

• I want to schedule an auto sanity check, so level of SOA sanity will be maintained over time
  • => UI to control web discovery recording done by dedicated filter