Let's Encrypt Twig Outage

[Matthew McPherrin]

Today, Let's Encrypt's "Twig" Log had an approximately 2 hour 18 minute outage on its write path.  

The monitoring endpoint remained available throughout.

This was caused by an outage in AWS's us-east-1 region of their DynamoDB product, which Twig uses for locking, to prevent issues like inconsistent STHes if multiple instances were incorrectly run.

Let's Encrypt runs our production logs (Oak, Sycamore, and Willow) in multiple AWS regions to mitigate exactly this type of issue. As a result, only Twig was affected. It recovered automatically once the AWS-side incident was resolved.