Might not be railway related, but we're seeing variances from 5ms to 150ms for a single HGET. There's not a ton of data in redis, but I'm not seeing this behavior when running locally so thinking it could potentially be railway related, whether it was a misconfiguration on my end.

We did migrate everything including redis to metal recently.

project: b6cc8a08-c46b-4418-8891-dcbc0557a83d