Hello Railway Team,

I'm experiencing a persistent discrepancy between Railway's reported Volume Usage and the actual filesystem usage inside my service.

Environment

• Plan: Hobby
• Application: Python Discord bot using SQLite
• Persistent volume: discord-recruit-bot-volume
• Mount path: /data

Measurements

September 30, 2026:

• Railway Volume Usage: 1.28 GB
• Filesystem usage (df): 980 KB
• SQLite database: 632 KB
• Cumulative process disk writes: 147.13 MB

October 27, 2026:

• Railway Volume Usage: 1.40 GB
• Filesystem usage (df): approximately 1.18 MB
• SQLite database: 864 KB
• Cumulative process disk writes: 1.34 GB
• SQLite WAL/SHM: Not present
• SQLite freelist pages: 0

Over approximately 27 days, Railway's reported Volume Usage increased by about 120 MB, while actual filesystem usage increased by only around 0.2 MB.

I understand that SQLite performs repeated writes to existing database pages, and that cumulative disk writes are not equivalent to allocated storage.

However, I would like to understand why Railway's reported volume usage continues to increase significantly while df, du, and the actual database size remain nearly unchanged.

Could you please clarify:

1. What exactly is included in Railway's Volume Usage metric?
2. Can repeated SQLite writes to existing database pages increase the reported Volume Usage without corresponding growth in df or du?
3. Is this discrepancy expected, or could it indicate an issue with storage allocation or usage accounting?
4. Is there a recommended way to diagnose the difference?

I have attached screenshots showing the 30-day Volume Usage graph and filesystem diagnostics.

Thank you for your help.

Note: The diagnostic screenshots are displayed in Korean because they were generated by my bot's custom monitoring command. The relevant measurements have been translated into English in the report above.

The screenshots are provided as supporting evidence of the actual filesystem usage and cumulative process disk writes.