Implement endpoint configuration Indicator and management

[dannycohen]

As Opie, I want to be notified of any configuration issues that may affect the accuracy of the ServicePulse display. I want to know when I need to update, reconfigure an existing endpoint, or stop monitoring an obsolete endpoint

Visualization:

1. In the dashboard, the configuration indicator is displayed indicating the status of the endpoint monitored by ServicePulse:
   • It is green when:
     • all monitored endpoints have a plugin installed
     • plugin version applies to minimal requirements (configured value of minimum plugin version supported by current version of ServicePulse)
   • It is red when;
     • one or more of the monitored endpoints do not have a plugin installed
     • the installed plugin is of an earlier version than the version specified as minimum requirement for the current ServicePulse version
2. Clicking on the configuration indicator in the dashboard leads Opie to the endpoint configuration page
   • Opie can view the endpoints list
   • Opie can un-check one or more endpoints, making the un-monitored
     • An un-monitored endpoint will remain visible in the list of endpoints
     • An un-monitored endpoint is not expected to send heartbeat messages, and failure to receive heartbeat messages from this endpoint does not result in the heartbeat indicator being red

Notes:

• An un-monitored endpoint can still be active (i.e. send custom checks events, failed messages and process SLA events; un-monitoring affects only the expectation for a periodic hearbeat message)

Demo / Acceptance tests:

Case 1:

1. Configure the Video Store sample to have the plugin installed only on the Sales and eCommerce endpoints (2 of 5 endpoints)
2. All 5 endpoints will be sending auditing and error data to the audit and error queues monitored by ServiceControl
3. Run the Video Store sample (5 endpoints)
4. The Configuration indicator in the dashboard page is green displays 2 monitored endpoints (after receiving heartbeat messages from the Sales and eCommerce endpoints)
5. 2 Configuration events are added (severity: info) indicating that 2 new endpoints were added to the monitored endpoints list (including endpoint names and hosting machine name and IP) and that heartbeat messages were received from these endpoints
6. Send a purchase order using the Video Store UI and await the successful completion of the purchase process
7. The Configuration indicator in the dashboard page turns red, and displays that 3 endpoints are not configured
   • (i.e. the 3 of 5 endpoints belonging to the Video Store sample that do not have a configured plugin; these endpoint were added to the endpoints list through the processing of auditing messages, and they expected by default to be monitored)
8. 3 Configuration events are added (severity: warning) indicating that 3 new endpoints were detected and were added to the monitored endpoints list (including endpoint names and hosting machine name and IP), but that they do not have a plugin installed and no heartbeat messages were received from these endpoints
9. Click on the Configuration indicator in the dashboard page and reach the endpoint configuration page
10. There are 5 endpoints in the endpoint list.
    • All are marked as monitored
    • 2 (sales and ecommerce) have a plugin installed and the plugin version is displayed
    • 3 have no plugin installed and there is no plugin version displayed
11. Uncheck the "CustomerRelations" endpoint
12. 1 Configuration event is added (severity: info) indicating that the "CustomerRelations" endpoint is no longer monitored
13. Return to the dashboard and make sure the configuration indicator in the dashboard page is red but that it now displays that 2 endpoints are not configured
14. Stop the Video Store sample and have the plugin installed on all 5 endpoints
15. Run the Video Store sample (5 endpoints)
16. The Configuration indicator in the dashboard page is green displays 4 monitored endpoints
17. 2 Configuration events are added (severity: info) indicating that 2 monitored endpoints (already listed in the endpoints list) started sending heartbeat messages
18. Click on the Configuration indicator in the dashboard page and reach the endpoint configuration page
19. There are 5 endpoints in the endpoint list.
20. Check the "CustomerRelations" endpoint to make it monitored again
21. 1 Configuration event is added (severity: info) indicating that the "CustomerRelations" endpoint is being monitored
22. Return to the dashboard and make sure the Configuration indicator in the dashboard page is green and displays 5 monitored endpoints

[indualagarsamy]

@dannycohen - Can we bump this out?

[dannycohen]

Lets discuss along with #13 (comment)

[dannycohen]

@andreasohlund / @johnsimons / @indualagarsamy / @sfarmar / @SimonCropp -

Following our discussion - the definition of the endpoint list behavior is in item 2 above. Specifically -

... leads Opie to the endpoint configuration page

• Opie can view the endpoints list
• Opie can un-check one or more endpoints, making them un-monitored
  • An un-monitored endpoint will remain visible in the list of endpoints
  • An un-monitored endpoint is not expected to send heartbeat messages, and failure to receive heartbeat messages from this endpoint does not result in the heartbeat indicator being red

Notes:

• An un-monitored endpoint can still be active (i.e. send custom checks events, failed messages and process SLA events; un-monitoring affects only the expectation for a periodic hearbeat message)

[dannycohen]

@johnsimons / @andreasohlund -

Following our discussion:

1. Configuration will be added as a separate side / menu link, and not as an indicator
   • Additional indicators in the future may use the information / functionality in the configuration section, but for v1 it will not be highlighted on the indicators bar
2. Configuration will show a list / table of identified endpoint instances and their properties
   • Endpoint properties will include endpoints identification properties and any additional information available on the endpoints for v1 (no specific requirements there).
   • The fields in the wireframe are nice-to-have and not mandatory
3. The Configuration page will include a "Monitor Heartbeats" checkbox for each endpoint instance
4. The "Monitor Heartbeats" checkbox will be checked by default
   • I.e. we will be using an opt-out policy for un-monitoring endpoints
   • All identified endpoints in SC will be considered as being monitored unless explicitly the user un-checks the "Monitor Heartbeats" checkbox.
5. The "Endpoints" indicators will be affected by the status of the "Monitor Heartbeats" checkbox for each endpoint:
   1. Total number of endpoints expected to send heartbeats = total number of monitored endpoints
      • Monitored endpoint = "Monitor Heartbeats" checkbox is checked
   2. A failure to send heartbeats by en endpoint who's "Monitor Heartbeats" checkbox is un-checked will not affect the displsy and will not be included in calculations of the Endpoints Indicator
   3. Endpoints who's "Monitor Heartbeats" checkbox is un-checked will not appear in the the "Endpoint Overview" page lists of endpoints (neither "active" nor "inactive" lists)
6. Note that the "Monitor Heartbeats" checkbox is applied to all identified endpoints.
   • Meaning, also to endpoints that never sent a heartbeat in the past.
   • For example,if an endpoint sent an audited message, and that endpoint has been identified - it will be added by default to the monitored endpoints list

Comments ?

[dannycohen]

// CC @indualagarsamy - see comment above.

[dannycohen]

@johnsimons - Seriously ?... open the Champagne !!! 👍

[johnsimons]

@dannycohen it works, but the ui needs a bit more love.
Let me know what you think, and I'll update it if needed ?