Sorry, ditching the template because I think I can explain it clearly enough and I'm going to submit a PR asap with a possible fix.

Basically as can be seen here https://github.com/mysqljs/mysql/blob/cf5d1e396a343ffa0fba23b3791d2dd5da20f30e/Readme.md#connection-options the documentation for the charset config option in mysql is (and has been so the last 4 years):

charset: The charset for the connection. This is called "collation" in the SQL-level of MySQL (like utf8_general_ci). If a SQL-level charset is specified (like utf8mb4) then the default collation for that charset is used. (Default: 'UTF8_GENERAL_CI')

But in the Knex table compiler code there's the following https://github.com/tgriesser/knex/blob/26dfeb8805bb11b63d76065d244c78c7b380b41b/src/dialects/mysql/schema/tablecompiler.js#L34-L41

which sets what may either be a charset (that mysql finds the default collation for) or a collation as the charset. This therefore fails all my migrations when I changed to using utf8mb4_unicode_ci as my charset. Also, as far as I can see mysql has never accepted a collate option, so I'm not sure where that comes from?

I have a fix in mind and will work on it right now and submit a PR asap, I just thought it would be good form to submit an issue first, then I can also link to this for the reasoning for my PR.