Mulitple On-Premise Agents with monitoring/failover
under review
D
Dave Manor
With only 1 On-Premise Agent (per documentation) the entire environment's on premise monitoring is dependent on a single agent; if that agent fails you lose monitoring of every other inspector that relies on it. Even if 2 on-premise agents are installed there is no way to specify it as a secondary inspector per environment (or per agent) or anyway to monitor and failover to it if other on-premise agent(s) fail.Recommend a redundant system for on-premise agents/inspectors so there is no single point of failure.
T
Ted Thueson
This is a must for cloud-only clients with in-office presence. Many clients are moving to serverless environments but still have on-premises network equipment we need to be able to monitor.
To expand on Dave's idea:
Option 1 - Network Rules on Inspectors:
Let us configure network rules on Network inspectors. For example, allow us to define the Public IP (WAN) the network device exists behind and automatically fail over to agents behind that same Public IP.
More specific example: SonicWall Inspector - add a launchpoint field to apply an "Agent Public IP Failover". If the assigned agent is offline, use Liongard's data for other agents behind the same Public IP to inspect from.
Option 2 - Smart Agent Selection:
Remove the need for agent assignment entirely and do it purely based on Public IP. The inspector should "elect" a viable agent that has the same Public IP under that same environment to run the inspection from.
In failure cases, it's easy to log that no viable agents are found.
Many, many (did I say many?) hours are spent managing the agents related to network inspectors. A huge benefit of Liongard is that we can apply any agent to these Network inspectors. We need to take this to the next level to make the platform more reliable for cloud-first environments.