Infrastructure

Monitor → Infrastructure shows every host that reports to your organization.

Views

Switch views at the top right. Your choice is remembered.

View	What it shows
Host map	One tile per host, colored by the selected metric from green to red. Spot hot hosts at a glance.
Rings	One ring gauge per host.
List	A table with status, OS, CPU, memory, fullest disk, files followed, agent version and last seen.

Fill by and group by

• Fill by picks the metric that colors the map: CPU %, Memory %, Disk % (fullest disk) or Swap %.
• Group by splits the map into groups: by any tag key (for example env, role or team), by OS, or by status.

Tags come from agent.tags in each host's logcore.yaml, so a consistent tagging scheme such as env, role and team makes the map much more useful.

Filters

• The search box matches the hostname, OS and tags, for example env:prod web.
• Side facets: Status (online or offline), OS, and one facet per tag key.

The summary bar shows online and offline counts and the average of the "fill by" metric for the hosts you're viewing.

Host page

Click a host to open its page:

• Overview: current CPU, memory, fullest disk, uptime and agent version.
• Charts: CPU, memory, disk, network traffic and load (1 min) over the selected range.
• Services: the log services configured on the host, with their files.
• Disks: usage per mount point.
• Host & agent: OS, platform, kernel, agent version, tags and last seen.
• Remove: deletes the host and its metrics. Logs stay until they expire. If the agent is still running, the host comes back the next time it reports.

Online and offline

The agent sends a heartbeat every 30 seconds. A host that hasn't reported for 2 minutes is shown as Offline. To get notified, create a host offline monitor.

Metrics collected

Metric	Details
CPU	Total utilization %
Memory	Used %, used and total bytes
Swap	Used %
Disk	Used % per mount point; the map uses the fullest disk
Load	1, 5 and 15 minute load averages (Linux)
Network	Bytes sent and received per second
Uptime	Since boot

They're sampled every 15 seconds by default (metrics.interval) and kept for 30 days.