Create clearly named profiles, connect the hosts you need, and inspect their metrics. Before opening a terminal, confirm the hostname and sample time. Saving several profiles does not mean every host is actively monitored.
Name hosts so environments are unmistakable
Names such as staging-api, prod-api and prod-db distinguish environments and roles. Keep each address, port and username with its profile; different ports on the same IP may identify different SSH services. These names are illustrative, not a record of tested servers.
What should you check when reviewing CPU and memory across multiple VPS hosts?
Check host identity, connection state and the latest sample time before comparing metrics. Hosts with different providers, capacities or roles cannot be ranked by a single utilization percentage.
- Hostname and role: distinguish staging, production and database hosts.
- Sample time and connection state: stale readings and disconnected hosts do not describe current usage.
- Memory and disk definitions: compare like with like and identify the filesystem being measured.
- Continue diagnosis: recheck hostname and time inside the selected host’s SSH terminal.
Verify one host at a time
- Verify the first host fingerprint and authentication, then check basic metric access.
- Repeat for the next host; a successful first connection says nothing about permissions on another.
- Separate network, authentication and metric-read failures before adding more hosts.
Compare timestamps and metric definitions first
Hosts may have different sample times, and an old value is not current status. CPU utilization, root-filesystem usage and network rates measure different things. Compare like metrics and check whether a disconnection stopped updates.
In shellivo, saved profiles, monitored hosts and open terminals are separate states. Overview and short-history features depend on the installed version. Use the currently published download as the availability reference, rather than assuming a development interface has shipped.
Open the terminal for the host you investigated
If CPU stays high, confirm the target host before reading its time and processes. The commands below apply to Linux with procps and show command names rather than full arguments. The ps %CPU statistic may differ from interval CPU shown by a monitoring chart.
hostname
uptime
ps -eo pid,comm,%cpu,%mem --sort=-%cpu | head -n 11Separate saved profiles from usage limits
Saving another profile does not increase an active-monitoring allowance. Check the current client and product page for limits and plan availability. Add a dedicated monitoring service when you need unattended alerts or shared historical data.
Questions
Do the servers need to connect to each other?
This workflow connects from your computer to each host separately. Your computer must reach each host’s SSH address and port.
Does saving a profile start collection?
Do not assume so. Check the actual connection state, sample time and current allowance for each host.
shellivo
Check supported platforms, available downloads and installation notes.