At 10.17 we were rearranging some of the newer servers to increase performance further. Part of this was changing the size of servers from many to fewer, but with much more resources. We are often changing the servers around to better serve you and your customers. However, this time we were going from many smaller servers to fewer larger, to be able to faster deploy hotfixes - if it would become necessary.
The issue we ran into was the number of threads each server could handle. This means that user's requests got queued, and customers experienced slower than normal response time both in rendering the recommendation boxes as well as when searching.
We are truly sorry for the inconvenience it had for you. We will keep monitoring the servers closely to ensure peak performance for you and your customers, especially on this important day.
If you have any further questions, don't hesitate to reach out to firstname.lastname@example.org.
Customers affected: all customers using our recommendations and search solutions.
Type of issue: Degraded performance
Timeframe: The issue was discovered 10.17 GMT+1 and solved 10.39 GMT+1.