AnsweredAssumed Answered

Portal health checking

Question asked by cadamsesriuk-esridist Employee on Feb 24, 2016
Latest reply on Jun 29, 2017 by HSompotan-esristaff

ArcGIS for Server has a health check URL so that it can be determined if there is an issue with the machine or one of the machines in an ArcGIS for Server site.

 

Is there an equivalent for Portal for ArcGIS? When Portal is configured for High Availability, the documentation just mentions the following. This is not a great way to know if there is an issue with one of the machines:

Check the status of a highly available portal

To check the status of a high availability configuration, use the command line utility and follow these steps:

  1. On any machine in the high availability configuration, log in with the Portal for ArcGIS account (selected in Step 2).
  2. Open a command prompt as an administrator (Run as administrator), browse to <Portal for ArcGIS installation directory>\tools\portalha, and run the portalha.bat tool with the -s command (for example,portalha.bat -s).
  3. Review the messages in the tool to obtain the current status of the high availability configuration.

The output is as follows:

 

C:\Program Files\ArcGIS\Portal\tools\portalha>portalha.bat -s

Machine '<machine.domain>' is configured for high availability architecture and includes the primary database.

The machine '<machine.domain>' is configured for high availability architecture and includes a standby database.

C:\Program Files\ArcGIS\Portal\tools\portalha>

 

A health check could be added to the web server by specifying a Portal URL such as the portaladmin site. Is there a better way to detect if a machine is not operational and prompt a failover?

 

A follow on question is that it doesn't seem possible to health check through a web adaptor. If one machine is down behind the web adaptor, the health check with still show the site as being successful. Do the health checks have to be directed through 6080/6443? If the web adaptor is being used in a reverse proxy configuration, how can you health check two endpoints through these internal ports?

Outcomes