shellivo · Shell · Guides · 中文

Check Linux CPU, Memory, Disk and Network over SSH

Inspect Linux resources with standard interfaces and common commands. Understand available memory, root-filesystem space and network byte counters without confusing totals with rates.

Updated · shellivo

After connecting to Linux, use /proc for CPU, memory and network counters, and df for filesystem space. Confirm the host and sample time first. Interval CPU utilization and network rates need two samples.

Confirm the server you are inspecting

These commands read information from an authorized Linux host. Replace the example username and address and verify the first-use fingerprint. Minimal systems may lack some common utilities.

ssh [email protected]
hostname
date -u
uptime

CPU: cumulative time is not current utilization

The cpu line in /proc/stat contains cumulative counters, not an instantaneous percentage. If a tool shows no value on its first sample, wait for another. Missing data is not 0% utilization.

head -n 1 /proc/stat
sleep 2
head -n 1 /proc/stat

Memory: look beyond free memory

MemAvailable estimates memory available to new applications without swapping. Caches affect how MemFree should be interpreted. Check definitions before comparing tools; low free memory alone does not establish exhaustion.

grep -E '^(MemTotal|MemFree|MemAvailable|Cached|SwapTotal|SwapFree):' /proc/meminfo

Disk: the root filesystem is not every mount

df -Pk / checks the root filesystem; df -Pk includes other mounted filesystems. Identify the mount that contains application data before investigating capacity. Root usage does not describe a separate data volume.

df -Pk /
df -Pk

Network: byte counters are not rates

Read the same interface twice and divide the receive or transmit byte difference by the actual elapsed seconds. After an interface reset or counter decrease, discard the old baseline.

Example: a receive counter increases by 2,097,152 bytes over 2 seconds, giving about 1 MiB/s. This illustrates arithmetic, not a product benchmark. Adding bridge and physical-interface traffic may count the same traffic twice.

cat /proc/net/dev
sleep 2
cat /proc/net/dev

Reduce repeated switching during inspection

Commands suit one-off verification. For repeated checks, shellivo can collect readings and provide a terminal for the corresponding host. When tool and command values differ, align sample times, interfaces and mount points before suspecting a collection problem.

Questions

Does a missing value mean zero resource use?

No. Check the connection, permissions, commands and sample interval to establish whether a valid new sample exists.

Do these commands replace continuous alerting?

No. They inspect status when run; continuous alerting also needs scheduled collection, storage, evaluation and notifications.

shellivo

Check supported platforms, available downloads and installation notes.

View downloads and system requirements

Keep reading

References