Is 1.7 Megs of Data Too Large for a Data Store if It Takes 20 Seconds to Execute?

Makes sense. Not the intended use case for data stores.

File storage would allow persisting larger files across runs.

/tmp would also satisfy within a run.

using an actual database would be the approach if you need to store larger amounts of data per entry.

Self service users are capped at 1mb of data store capacity on the Advanced plan. It is intended for small key value pairs.