Scenario

Run: opensre tests synthetic --scenario 003-storage-full

The RDS instance ran out of storage during a bulk INSERT archival job. FreeStorageSpace collapsed from 15 GB to 0 in 15 minutes.

Correct Answer

• Root Cause Category: resource_exhaustion
• Root Cause: Bulk INSERT job exhausted storage. FreeStorageSpace hit 0, writes blocked.

Human Review Checklist

• Does the agent identify storage exhaustion as the root cause?
• Does the agent reference the FreeStorageSpace metric trend?
• Does the agent cite the RDS event confirming "ran out of storage space"?
• Does the causal chain flow: bulk INSERT → WriteIOPS spike → storage fills → writes blocked?
• Does the agent note WriteIOPS collapsing to zero (signature of full disk)?

What Good Output Looks Like

Agent traces FreeStorageSpace from 15 GB → 0 during high WriteIOPS, cites the RDS event as definitive confirmation, and explains the bulk INSERT as the driver.

Common Failure Modes

• Agent mentions storage is low but doesn't explain what caused it (misses the bulk INSERT)
• Agent ignores the RDS event (definitive evidence) and relies only on metric trends
• Agent doesn't notice WriteIOPS collapsing to zero (key signal that writes are physically blocked)