Logs are being fully ingested again.

The issue was with the internal workers in our shared OpenTelemetry service. They became unhealthy but weren't rolled. We're looking into why and how to prevent this from happening again.