Bash script should only kill those instances of another script’s that it has launched

Run pgrep -f "ssh.*-D" and see if that returns the correct process ID. If it does, simply change pgrep to pkill and keep the same options and pattern

Also, you shouldn't use kill -9 aka SIGKILL unless absolutely necessary because programs can't trap SIGKILL to clean up after themselves before they exit. I only use kill -9 after first trying -1 -2 and -3.

You launch foo.sh as . ./foo.sh.

A shell script is not a process by itself. It is interpreted by an instance of the shell, which is the process.

. is an alias of the source command that executes commands in the current shell environment. This means ./foo.sh is executed by the same shell that executes the script that launches it. It does not create a new process and this is why you cannot find it using ps.

You can launch foo.sh as ./foo.sh. In this case it will be executed by a new instance of the shell, i.e. into a new process.

But, depending on what it does, the execution of the original shell after ./foo.sh completes might be different in the two cases. When foo.sh is executed in the same shell, the environment variables it changes belong to the shell that also executes the original script. Changing them could affect the original script.

When you run foo.sh in a separate script the changes it operates on the environment variable do not affect the original script. They are different processes, the do not share anything.

but how can I check if foo.sh is running at all?

This is easy from the script that launched foo.sh because it waits until foo.sh completes before resuming its execution.

It is probably impossible (or almost there) from outside.