Our server's main NVME drives started failing due to old age and read-write wear, causing our services to intermittently crash. This has been an issue for the past few weeks, and we only figured out the actual root cause very recently. We're working on installing more drives to mitigate this issue.
We expect to resolve this problem definitively by the weekend. In the meantime, our services may be unstable and be intermittently unavailable as we continue our maintenance work.