After further investigation, the performance issues were limited to a small subset of hosts including two customer cluster hosts in sa-east-1 and one customer cluster host in us-east-1. The affected hosts have been rebooted and no further load issues have been observed.
Posted 19 days ago. Jul 02, 2019 - 18:36 UTC
We are currently investigating an issue causing higher than expected load on multiple hosts in multiple regions following a recent release. This could result in degraded customer performance and difficulty in executing plan changes, though existing clusters remain accessible.