MongoDB is not the only document database in existence (for example, ElasticSearch, RavenDB to name a couple off the top of my head). I would first consider what you need out of the JSON data.
- Efficient searching?
- Ad hoc queries?
- Faceted searches?
- Geospatial queries?
- Histogram results?
- Risk of data corruption or loss? (This used to be a serious problem with Mongo, but not sure if it still is)
When you examine what your actual needs are with that data you may choose something other than the two options you listed (MongoDB and MySQL/MariaDB). You need to know that your solution can handle the data needs now, and can scale in the future. You also need to understand how that scaling works.
That said, you can go with a short-term/long-term solution. For example: start with MySQL JSON data now since you already have that infrastructure, and start evaluating alternatives for the future. When your queries get too inefficient for MySQL, you should know very well the tool you want to use.