Skip to content

Agents & collectors

Vesana has two ways to gather data:

Component Where What
Agent On the monitored machine Local checks: CPU, RAM, disk, service status, logs, custom scripts
Collector Linux VM in customer network Remote checks: SNMP, ping, SSH, HTTP, discovery

Both push-based — no inbound ports on monitored devices, no inbound from server to customer network.

When agent

  • Linux/Windows servers where you can install software
  • Workstations
  • Container hosts (Docker, Podman)
  • Anywhere local data is needed (service status, event log, per-mount disk)

When collector

  • Switches, routers, firewalls, UPSes, storage — anything reachable only via SNMP/ping
  • „We can't install anything there"
  • Discovery scans
  • When monitoring an isolated network (collector is the only outbound talker)

When both

Combined is fine and common:

  • Linux server: agent for local stuff, collector for ping (sees if routing works)
  • Storage: collector for SNMP, agent (on a server nearby) for pings

Architecture comparison

flowchart TB
    subgraph "Customer network"
        S1[Server with agent]
        S2[Server with agent]
        SW[Switch via SNMP]
        UPS[UPS via SNMP]
        C[Collector VM]
        SW -.SNMP.-> C
        UPS -.SNMP.-> C
    end
    subgraph "Vesana server"
        R[Receiver]
    end
    S1 -->|HTTPS POST| R
    S2 -->|HTTPS POST| R
    C -->|HTTPS POST| R

In this section

  • Agent (Linux/Windows)

    Installation, configuration, systemd / Windows service, token rotation, troubleshooting.

  • Collector

    One-command installer, distros, discovery prerequisites.

  • Agent versioning

    Auto-update mechanism, VERSION file, rollout pause, NSIS wizard.