Avoid duplicate/redundant data-shift reports

[thesoftwarephilosopher]

Essentially, we're sometimes seeing the same data-shift errors being reported several times in a row, when once would be enough.

I'm not sure what the right solution is right now. Putting this here so we can prioritize it and have a centralized place to discuss it.

But my gut feeling is that each type of data-shift error needs its own way of checking:

1. "Late transactions/licenses" needs to check against all past data sets, but only look in the current data set for new appearances
2. "Deleted transactions/licenses" also doesn't need to check every past data set against every data set before it, only the current one
3. I haven't thought of the other ones, as these are the only types we've dealt with so far for this issue

[bberenberg]

Why not take the "error message", hash the raw text, store the hash, don't print messages with stored hashes?

[thesoftwarephilosopher]

Because that would make too much sense 🤦

[boris-moduscreate]

Prob an env variable here where we set the max number of times a given error is posted? Default 1
And a second option where all records are wiped out after a successful run? Default disabled