We have regular backups and redundancy setup on all customer data services (language model, analytics).

We have processes to leverage Azure support for infrastructure (non-code related) issues and use application logging and alerting to be aware of incidents and help with tracking the root cause.

We leverage infrastructure as code to deploy our infrastructure and have designed our systems to be able to reset the infrastructure quickly and restore backed-up data.