I have a very low volume API (c. 20 hits per hour) fronted by Cloudflare. The primary purpose of Cloudflare is to provide a IPv4-to-IPv6 proxy.
This has run trouble-free for over 12 months until between 9th and 10th August where ever since there have been intermittent timeout errors.
Normally this API is called from AWS Lambda function. From AWS monitoring is where I’ve been able to determine the start date. See 6 week history below.
When trying to connect to API from home computer I’ve managed to re-create the issue. On first connection, it sometimes times out. If I then refresh the page, the API works immediately.
When connecting to origin server direct I’ve not managed to trigger a timeout.
It’s like there has been a configuration/software change at Cloudflare around the 10th August that has caused this effect.
This causes a problem as Cloudflare connection is unreliable.