For each runnable providing health information, a different set of checks is performed. As an example, the set performed for the abs runnable is described here:
This check compares the average time it takes to access a synchronized resource with the average time it takes to access a non-synchronized resource on the system. From the measured times it calculates a ratio. A high ratio means that access to synchronized resources is relatively slow, which could have a negative effect on the computation speed of ARIS.
The check fails if the ratio is greater than 30. Because a large set of test data objects is necessary to accurately calculate the result, the check is executed only once an hour in the background.
This check detects threads running longer than 5 minutes. This could indicate an infinite loop. All AJP and HTTP threads currently running are constantly monitored.
This check detects whether the JIT compiler is active. If the JIT compiler is active, the check also monitors the state of the used code cache volume.
The check fails if the JIT compiler is not active or if the used code cache volume is at 90% of the maximum.
This check helps to identify JIT and VM performance problems. The check executes a short but extensive computation and measures the time needed to calculate the result.
The check fails if the calculation takes more than 2,5 seconds.
This check detects whether the number of documents (revisions included) stored in ARIS document storage complies with the license in use.
Green: If the number of documents is < 40.000.
Yellow: If the number of documents is >= 40.000.
Red: The check fails if the number of documents is > 10.000.000.
This check identifies the host name of the machine via its IP address.
The check fails if the host name cannot be resolved or if the call takes longer than 30 milliseconds.
Checks whether the server is connected to ZooKeeper. If the server is not connected, the check fails.
Ensures that there is a leading abs runnable instance for the abs application type registered in ZooKeeper.
Checks whether a database connection can be created and measures the time for executing a set of standard database queries.
The test fails if the database connection cannot be created or if the average execution time is longer than a calculated threshold.
This check tests whether a valid database connection can be created and measures the latency between the ARIS system and the database backend via multiple ping requests.
The check fails if the database cannot be reached or if the average response time exceeds 300 milliseconds.
The check detects the number of active database connections and the maximum number of active database connections in the system.
The check fails if the number of active connections is equal to or exceeds 95% of the maximum.
This check obtains all registered CloudSearch instances in the system and sends a ping request to them.
The check fails if there are no CloudSearch instances registered or if no CloudSearch can be reached.