Task: Create Batch Job to Run Aggregation Query and Update `hit_count` records

[nelsonic]

As noted in #355 the aggregation query (SELECT COUNT(*) FROM hits WHERE repo_id = 123) 💭
takes 2 MINUTES on a table with 30 Million Rows. ⏳
And is the reason the request/query time-out and crashes the server! 💔

So we're implementing hit_count 🆕
which is just an integer counter 1️⃣
to return millisecond responses again. 🏎️ 🏁 🎉

The problem is that we are temporarily ignoring the historical data in the hits table ... 🙄
But the data is not "lost" it's just in stasis. 💭
So ...

Todo

• Create a batch job to go through the hit_count table
  and update count to reflect the aggregate from hits. 🧑‍💻
• This must be done gradually to avoid crashing the server again. ⚖️
• Once all records are updated, we can sunset this data migration script. 🌆
• I expect writing the script to take around T1h and running it to take T1w ⏳
  (Running the script once per minute to process a single record. 1440 runs per day. 10,000/week)

Will work on this immediately after I have #355 working.

[nelsonic]

A Practical Guide to Elixir’s Task Module — Let’s create a Cron Job: https://medium.com/@Vickris/a-practical-guide-to-elixirs-task-module-let-s-create-a-cron-job-4c6afe74b846