Monitoring

Platform provides an health check mechanism which can be invoked by external IT monitoring systems in order to check out if the current node is working correctly.

This web service is public and can be invoked as a Rest Service. In the easiest case, the invokation URL is

https://<domain>/platform/healthCheck

If the node is working, this service will respond with an HTTP 200 status and no content.

There are additional settings you can pass forward as request parameters, described as follows.

OK output

https://<domain>/platform/healthCheck?gimmeOk=true

If the node is working, this service will respond with an HTTP 200 status and content set to "OK".

Database status

https://<domain>/platform/healthCheck?checkBaseDatasourceToo=true

If the node is working, this service will respond with an HTTP 200 status and content set to something like:

Time required to execute a query on the main repository using EntityManager: 28ms
Time required to get a connection on the main repository: 0ms
Time required to execute a query on the main repository using JDBC: 1ms

Consumed memory

http://localhost:8280/wag/healthCheck/traceUserSessions?json=true

If the node is working, this service will respond with an HTTP 200 status and content set to something like:

{
"totalMemory": 210,
"freeMemory": 150,
"consumedMemory": 60,
"currentActiveSessions": 0,
"sessions": [

]
}

Out of memory errors

https://<domain>/platform/healthCheck?checkMemory=<minutes>

This web service will check if there have been out of memory errors in the last <minutes>.

This call is more helpful than the first one, since an out of memory could still allow a web application to run and respond with an HTTP 200, but this does not mean that the web app is in a stable state: checking for out of memories can avoid unpredictable behaviors by the application and inform the calling system to restart the VM.

In case of out of memory errors found, an HTTP 500 error is returned.