Skip to content
WatchTower15 Docs

Metrics & telemetry

Everything on this page is collected with zero privilege — the agent reads standard, world-readable /proc and /sys paths through the gopsutil library. No metric here requires root or a Linux capability.

There are two collection cadences: a lightweight telemetry push every 60 seconds, and a heavier system-information snapshot every 300 seconds. Both intervals are profile-configurable.

Telemetry — every 60 seconds

Section titled “Telemetry — every 60 seconds”

A compact, high-frequency vitals sample for time-series charts and metric alarms.

Field Source Description
cpu_usage_percent /proc/stat Aggregate CPU utilisation (1-second sample)
memory_used_mb /proc/meminfo Memory in use
memory_free_mb /proc/meminfo Unallocated memory
memory_cached_mb /proc/meminfo Kernel page cache
swap_used_mb /proc/swaps Swap in use
disk_used_gb statfs(/) Root partition used
disk_free_gb statfs(/) Root partition available
disk_io_read_mbps /proc/diskstats Block-device read throughput (1-second delta)
disk_io_write_mbps /proc/diskstats Block-device write throughput (1-second delta)
net_in_mbps /proc/net/dev All-interface inbound throughput (1-second delta)
net_out_mbps /proc/net/dev All-interface outbound throughput (1-second delta)
load_avg_1m /proc/loadavg 1-minute load average
load_avg_5m /proc/loadavg 5-minute load average
load_avg_15m /proc/loadavg 15-minute load average
process_count /proc Total running processes

System information — every 300 seconds

Section titled “System information — every 300 seconds”

A fuller inventory snapshot, split into a host record plus per-disk and per-interface breakdowns.

Host snapshot

Section titled “Host snapshot”

os_name, os_version, kernel_version, hostname, cpu_model, cpu_cores, total_memory_mb, uptime_seconds.

Per-disk (one row per mounted partition)

Section titled “Per-disk (one row per mounted partition)”

device, mount_point, filesystem_type, total_gb, used_gb, free_gb, inodes_total, inodes_used.

Per-interface (one row per non-loopback interface)

Section titled “Per-interface (one row per non-loopback interface)”

interface_name, mac_address, ip_addresses[], rx_bytes_per_sec, tx_bytes_per_sec, is_up.

What requires privilege (and is therefore not collected by default)

Section titled “What requires privilege (and is therefore not collected by default)”

To keep the guarantee honest, here is what the default zero-privilege agent deliberately does not collect, because it would require elevating the agent:

Capability Would require
Per-process inspection of other users’ processes CAP_SYS_PTRACE
Raw packet / flow capture CAP_NET_RAW + CAP_NET_ADMIN
Filesystem integrity monitoring root or CAP_DAC_READ_SEARCH
Security/auth-log access on some distributions root or a distro group

Where these become available they are delivered by separate, scoped helper binaries that each hold only the one capability they need — never by widening the main agent. See the privilege model.

Docker container metrics (opt-in)

Section titled “Docker container metrics (opt-in)”

If you enable the Docker metrics flag at install, the agent additionally reports running-container counts and per-container CPU/memory/network stats. This requires docker group membership, which carries the root-equivalent caveat noted on the installation page.

Section titled “Related”