Error rate is back to zero and service is fully restored.

Due to a Redis cluster node becoming saturated Function input fetching and output publishing has an elevated failure rate.

We're actively working on remediation to bring down error rates to zero.