Cannot reuse same --group name across parallel machines

I am running my tests in parallelisation.
`parallelism: 2`
`cypress-repeat -n 3 --until-passes`
CI Provider: CircleCI

`Machine#1`: Test fails --> cypress-repeat reruns with `--group=run-1-of-3` (all good 👍🏽 )

`Machine#2`: Test fails --> cypress-repeat reruns with `--group=run-1-of-3` (fails 👎🏽 )
But, this group name `run-1-of-3` is already used up in Machine#1, hence cypress run fails.

Workaround: Can we also include the Machine# , while creating the groupname?
Eg. Circle CI parallel machines have unique index in them, identified env variable `CIRCLE_NODE_INDEX`. [more info](https://circleci.com/docs/2.0/env-vars/)

We can form the group name like so: `${group}-${CIRCLE_NODE_INDEX}-${k + 1}-of-${repeatNtimes}`

I don't know how parallel machines are identified in other CI providers.

Below is the stdout of the CI failure from `Machine#2` (`Machine#1` ran all fine)


Start by tracing how cypress-repeat constructs the --group value for repeated runs, then inspect the CircleCI CIRCLE_NODE_INDEX environment variable described in the issue. Done means parallel machines produce distinct group names while existing non-parallel and non-CircleCI runs continue to work.

Written by the indexing model from the issue text.