Discrepancy between Volume Usage Dashboard (~3GB) and actual disk usage (52MB)

Question

I am seeing a significant discrepancy between the **Volume Usage** reported in the Railway dashboard and the actual usage reported by the OS inside the container.

* **Observed Usage (Dashboard):** Consistent growth up to **\~3.00 GB**.
* **Actual Usage (Container CLI):52 MB** (verified via  
```  
df -h /data  
```  
and  
```  
du -sh /data  
```  
).

**Environment Details:**

* **Mount Path:**  
```  
/data  
```
* **Filesystem Type:** zfs (detected via device  
```  
/dev/zd37440  
```  
)
* **Database:** SQLite 3 (primary file is \~54MB).

**Diagnostics Performed:**

1. Ran  
```  
du -sh /data  
```  
which returns **52M**.
2. Ran  
```  
df -h /data  
```  
which returns **52M Used** and **4.5G Available**.
3. Verified directory contents via  
```  
ls -laR /data  
```  
. No large hidden files or journal logs (WAL/SHM) account for the missing \~2.9GB.
4. No "open deleted files" appear to be inflating the block usage.

Could you please clarify if this 3GB represents provisioned blocks, snapshot overhead, or if there is a known delay in the volume metric aggregation? I want to ensure I am not actually nearing the 5GB volume limit while the container reports only 2% usage.