How to ensure complete data deletion of decomissioned tenants from local TSDB (not using object storage system yet)

[nicktill]

I'm using Thanos Receive without an object storage system, and I need guidance on ensuring complete data deletion for a decommissioned tenant from the local TSDB located at /var/thanos/receive

While I understand the decommissioning process involves flushing in-memory samples to disk, I would like to understand detailed steps or best practices for identifying and safely removing specific data files from the TSDB path.

Here is an example of the file system structure:

/var/thanos/receive $ ls
ISV+Tenant2                     ISV+Tenant3                ISV+Tenant4           default-tenant
/var/thanos/receive/ISV+Tenant2 $ cd ISV+Tenant2
/var/thanos/receive/ISV+Tenant2 $ ls
01HPHPKJ91EQZCESDBPR9J2ADF  01HPJB6SJ4V4FRWBXH2WYS1A8N  01HPJZSXPHZQ0AYCJW3F3N1BZ3  01HPKMD2W2MQVJJ1W2YWRSX8MT  01HPM906NXYG1GE0RH8MRDT7DK  lock
01HPHXF794S8C3JRDCSDT4TT9Y  01HPJJ2C3KQCFPCQKFM467SD22  01HPK6NNYKRJQ34WF7KJ3S94QR  01HPKV8VKQR2700TXM1YVYYFKS  01HPMFVXDQCTZX074MCYSKE366  thanos.shipper.json
01HPJ4AZRXQJ9MJEM3H5PDFDYB  01HPJRY58C9H8JV4K5BMXMDGCW  01HPKDHA5XSVXB4SK0PE38AK1F  01HPM24MKCP1FGBCSM87W0J0XC  chunks_head                 wal

Our team initially had some scripts to remove tenant data by going into each tenant and deleting specific directories within them.
However the script was not functioning properly, as metrics would still show up for a tenant after deletion.

Based on this behavior, I would like to clarify if I simply can remove the tenant folder entirely if I want to remove data associated with a specific tenant, or if there is a better solution to accomplishing what was explained above? Any guidance/clarifications anyone might have would be greatly appreciated!