At about 17:33 UTC, the Objects service began to report errors and experience elevated latencies in the EU PoP. We started to implement rate limits on the impacting Objects query, during that process the latency resolved itself by 19:10 UTC. The reason for the issue is due to a lack of monitoring to detect runaway query processes within the Objects service.
To prevent a similar issue from occurring we have implemented the necessary monitoring to alert our on-call staff and established default rate limits to prevent incorrect queries from consuming resources unnecessarily.