Some services are down

Services have entered backfill and recovery mode as we are now able to access our storage.

Lightning recovered.

This was resolved.

We are considering this resolved at this time. Systems are running nominally and our error rate has reduced to less than 1%.
Apologies for the unscheduled downtime. Frontend systems were battling with the backend database and got caught in an infinite loop of trying to correct themselves. We've made changes to mitigate this in the future.