ccie18643 / netwolf Goto Github PK

NetWolf is a proof of concept for the distributed real time network monitoring and automation system capable of simultaneously interacting with thousands of network devices using vertical and horizontal scaling of resources. It consists of the manager program and number of agents. Agents can be spread out over multiple machines and bind separate processes to every available CPU core. Each of those processes maintains SSH connectivity to multiple network devices. To achieve high number of connections per core and the same time consume least amount of memory possible agent is internally based on Asyncio. Agent registers with manager and receives jobs that contain the hostname for the network device it's supposed to connect to and the set of commands that need execute on it. For the purpose of monitoring job can also contain filters that need to be applied to command output before its sent back to the manager.