Prometheus deployment is using a built-in snapshot feature of TSDB admin: https://prometheus.io/docs/prometheus/latest/querying/api/#snapshot. The resulting file with metrics is then copied using custom filesystem boshrealease to the backups bucket.
Why the default configuration specifies the
skip_head=true? Head of the TSDB consists of data which is partially in memory and partially on disk. This is not consistent and backing it up may fail: