There are 4 key components to a successful monitoring scheme for Forum Sentry.
1. Monitoring the device via SNMP Polling - This includes using a third party tool to periodically poll the devices via SNMP. This allows for collecting stats on the system resources (CPU, Memory Usage, etc..). Overtime this will build a "resource profile" so you know what percentages are in the norm, when the device is under particularly heavy load, etc. The MIBS required for this are included on the SNMP Monitoring screen in the WebAdmin interface. Instructions for configuring this are in the attached Monitoring and Reporting Guide.
2. Real Time Alerting - The Sentry IDP Rules can alert via email, SNMP trap, and logs. The Sentry runtime log messages can be sent to a remote Syslog server, where specific filters and alerts can be configured. A common scenario is to write all log messages off to a syslog server such as Splunk and then have rules in Splunk to alert on specific events.
3. Load Balancer Health Checks - Rather than having the load balancer simply check for a tcp connection to the Sentry policy, the load balancer should make an HTTP request and expect a certain response. You can build specific Sentry policies to receive the health check calls from the load balancer and return a valid response - or - Sentry can forward the healthcheck request to the backend server and proxy the response. Typically customers build "service mode" policies in Sentry that respond to the health check calls with a static doc and 200 OK status.
4. Automated Regression Testing - Using SOAPSonar you can build a comprehensive test suite for all policies in Sentry. These tests can be automated for complete Client>>Sentry>>Backend testing. These SOAPSonar projects (ssp files) can be scheduled to run automatically and email you the results. If anything in the results change, you'll be made aware of this in the report.
Please contact Forum Systems Support or open a new Helpdesk ticket if you have any questions.