Egress to M.A.R.C.U.S. Cloud
This is the only external firewall rule required for the Collector. Every connection uses TLS 1.2 or higher.
Outbound proxies. If your network sends outbound traffic through an explicit proxy, add
HTTPS_PROXY=http://[proxy-host]:[port] (and NO_PROXY for anything that must go direct) to the Collector’s env file: /etc/av-bridge/env on Linux, C:\ProgramData\av-bridge\.env on Windows. Then restart the service. Every Collector connection then uses the proxy: telemetry, commands, configuration, remote access and updates. The proxy must allow CONNECT to port 443, and WebSocket upgrades for remote access. Collectors older than v1.0.82 only used the proxy for remote access and updates, so update them first. The Collector’s requests, and the handshake that opens the remote access tunnel, are signed with an HMAC-SHA256 key unique to the Collector, so they are authenticated independently of the transport.*.involvecloud.com. Through the tunnel, the Collector only connects to its own devices, on ports 80, 443, 8080 or 8443, the device’s own control port, or the port in its web_port tag, never to anything else on the network.
Browsers used for remote access
Each remote access session has its own address, such as
<session>.access.involvecloud.com. That is one label deeper than *.involvecloud.com, and many web proxies and allow-lists match a wildcard to one level only, so allow *.access.involvecloud.com explicitly.
If your allow-list needs specific hostnames rather than a wildcard, Involve will give you the exact hostnames for your tenant on request.
Collector to devices
For an on-premises Collector, this is ordinary LAN traffic. A cloud-hosted Collector needs a routed path from your cloud tenant into each site’s AV VLAN.
Inbound to the Collector
You don’t need to open any inbound ports. The Collector only makes outbound connections, so there’s nothing to allow from the internet or from your own network.Optional: the Collector’s local API
The Collector also has a small local API on TCP port 8080, for health checks, Prometheus metrics, and device status and commands. The cloud doesn’t use it. Out of the box it listens on127.0.0.1:8080, so only the Collector host itself can reach it.
Only change this if another machine on your network needs it, such as a monitoring server. In that case:
-
In
config.yaml, open the API to the network and give it at least one key: - Restart the Collector.
- On the host firewall, allow port 8080 only from the machines that need it.
Authorization: Bearer [key]. The health check, /healthz, never needs a key, so monitoring keeps working.
A key is required. If you set a network address without a key, the Collector won’t open the API to the network. It logs an error and keeps listening on
127.0.0.1 only.Updated 06 Nov 2026 14:15 · page 369ca55