The situation is now stable and the service is answering without errors anymore.
We’re closing this incident.

We have applied a mitigation which brought back the service latency to a nominal state.

We are investigating increased latencies and error rates on the Object storage service. We’ll post an update as soon as we have more information.