CID/POD column: truncated names and container aggregation with same suffix

[dvargek]

Hi there,

when running atop with Docker container monitoring enabled, we noticed two problems related to the CID/POD column:

1. Truncated container names
Container names are cut down to 15 characters (UTSLEN = 15).
In our environment we use long, FQDN-like container names, e.g.:

lb2.lr-cl2-build.exampledomain
test4.lr-cl2-build.exampledomain

Atop only shows the last 15 characters, e.g. d.exampledomain - this makes it impossible to distinguish containers by name.

2. Containers with same suffix aggregated
Both lb2.lr-cl2-build.exampledomain and test4.lr-cl2-build.exampledomain appear as a single entry (d.exampledomain) in atop.

Resource usage is summed, and one of the containers is effectively hidden in the display/history. This is very misleading for monitoring and debugging.

Expected behavior

• The CID/POD column should display unique identifiers for each container (full name or unique ID).
• Each container should have its own row in atop, even if their names share the same suffix.

Actual behavior

• Column width is limited to 15 characters, causing truncation.
• If two containers end with the same 15 characters, they are merged into one line in atop.

We experimented with a small patch (using the help of AI) to increase UTSLEN and adjust the display logic. While this helped in our setup, we are not sure about potential side effects on data structures, backward compatibility of raw log files, or other functionality. We would prefer an upstream-supported solution.

Thanks for your work!