Advice on preventing race conditions caused by multiple calls to the same endpoint made simultaneously (using Redis)?

[brianjenkins94]

I have a somewhat long-running route that gets paginated data from a third party and uses apicache to cache the result of it. After adding Redis, I've noticed that I don't reliably get back the full dataset from the cache.

It's as if the data that's being cached is being streamed to the database, and the call that gets there second sees that there's one page of data in the cache and uses that as if it were the full response.

I'm working on a minimal sample, but wanted to know if this sounded plausible.

[brianjenkins94]

Naturally I found a bunch of additional information after submitting an issue:

• Caching happens only on the 2nd request to the endpoint arthurfranca/apicache-plus#26
• Support for concurrent requests? arthurfranca/apicache-plus#23
• The cache should give out "leases" if same URL is requested concurrently #33