On the morning of Sunday, December 14, 2025 our automated systems detected high latency on our provider search endpoints. Our on-call team investigated and determined that the latency was the result of a higher-than-usual write throughput on our search cluster and a misconfiguration of that cluster’s replica count.
To mitigate the issue, our team temporarily throttled writes to the cluster, allowing the provider search endpoints to return to their normal latency a little after 10am ET. We are in process with a longer-term fix to migrate the cluster to a new configuration and expect to complete that migration by Tuesday December 16, 2025. In the interim, we expect performance of our provider search endpoints to be at normal levels.
We apologize for this performance degradation and are committed to preventing future incidents going forward.