Monitoring Your Forwarders

Configure → Log Forwarders lists every forwarder you’ve installed, with its status, group, and recent activity.

Reading the status

Each forwarder shows a status dot:

Status	Meaning
Pending approval (greyed out)	A new forwarder has connected but hasn’t been approved yet. Approve it to enable delivery over an encrypted connection.
Online / healthy (green)	Reporting in and running normally.
Not seen yet / stale (amber)	Approved, but hasn’t checked in recently. Common right after install, or if the host has been offline.
Offline / unhealthy (red)	Was reporting, but has stopped.

Note

The status dot reflects whether the forwarder itself is checking in — it isn’t a direct measure of whether logs are actually arriving. A forwarder can sit at “not seen yet” for a while and still be shipping large volumes of logs. Use the Logs (past hour) count and sparkline on its row as the real signal for delivery, and confirm in Log Search.

Confirming delivery

Open a forwarder’s row for its detail drawer:

• Logs (past hour) and a sparkline of recent volume — the clearest signal that data is actually flowing.
• Config — whether its current configuration (from its assigned Group’s Profiles) has been applied.
• Last seen and its reported version.

For a definitive check, go to Log Search and filter to the host — you should see recent entries.

The host’s own metrics

A forwarder’s Metrics tab charts the machine it runs on, over the last hour, 6 hours, or 24 hours:

• System — CPU, memory, and disk read/write. These need no configuration; every connected forwarder reports them.
• GPUs — a card per NVIDIA GPU in the host, with utilization, memory, temperature, and power draw. This one is a module you turn on for the host’s profile, and it is Linux only — see Turning on GPU metrics.

Seat usage

Log Forwarders share a plan-wide seat allowance. Configure → Log Forwarders shows how many you’re using against your plan’s limit. If you’re at your limit, approve fewer forwarders or upgrade your plan — see Pricing.

First steps when a forwarder goes quiet

1. Confirm the process is running on the host:
   • Linux: sudo gridnms-forwarder status or sudo systemctl status gridnms-forwarder
   • macOS: sudo gridnms-forwarder status
   • Windows: gridnms-forwarder status or Get-Service GridNMSForwarder
2. Check the agent’s own log for errors — the exact path is in Installing a Log Forwarder for each platform.
3. Check network reachability from the host to the collector it’s configured to use — the forwarder only makes outbound connections, so a firewall or routing change on either end can cut it off.
4. Check the collector itself — if every forwarder pointed at one collector goes quiet together, the problem is the collector, not the forwarders. See Monitoring & Managing Collectors.
5. Once connectivity returns, a forwarder reconnects on its own — nothing to re-approve, and anything queued while it was disconnected delivers automatically.

Tip

Most forwarder problems are connectivity, not something wrong with the forwarder itself. Start with “can this host reach its collector?” before digging further.

Related pages

• Forwarder Management — Groups, Profiles, and what each forwarder collects.
• Monitoring & Managing Collectors — the collector side of the same connection.
• Logs & Log Search — confirm what’s actually arriving.