We know that WDQS is getting close to the scaling limits. While we want to make WDQS scale, we need to have a plan if we reach a significant road block and we’re not able to support the system at all.

Random notes that need to be organized into a concrete plan:

• can we anticipate the failure modes?
• list worst case scenarios
• potential degraded modes
• investigation of queries to understand impact of various degraded modes
• we need to address both the technical side and the “business” side (how to influence our users and what they do, what we want to / can support)