Load has been reduced on the server, and it appears to have stabilized. We are continuing to monitor the performance.
Posted Oct 25, 2020 - 18:24 UTC
The CloudAMQP log system is sporadically unresponsive. This is preventing messages from being shoveled from the logstream queue on the root vhost of customer's RabbitMQ brokers. Customers with dedicated instances might see buildup of messages in that queue and may not see current log entries on the CloudAMQP console.