Considerations: flatfiles (currently using YAML) aren't a good solution in production.
Cache will eventually get quite large as more repos are added, so a full-fledged DB is likely to be useful.
Some small calculations (like calculated user ranking scores) might be appropriate to store in something like Redis.
It would likely be useful to invalidate cache in a very granular way, e.g. for particular issues, without invalidating the entire cache for a repo.
Considerations: flatfiles (currently using YAML) aren't a good solution in production.
Cache will eventually get quite large as more repos are added, so a full-fledged DB is likely to be useful.
Some small calculations (like calculated user ranking scores) might be appropriate to store in something like Redis.
It would likely be useful to invalidate cache in a very granular way, e.g. for particular issues, without invalidating the entire cache for a repo.