diagnose

Server Diagnostics

The diagnostics script is at tools/diagnose.sh. It provides safe, read-only commands for inspecting the remote server memory at /home/ec2-user/memory.

Commands

Status (overview)

Get overall system health:

./tools/diagnose.sh status

Shows: container status, memory, disk, and recent errors.

Logs

View docker logs:

./tools/diagnose.sh logs              # all services, last 100 lines
./tools/diagnose.sh logs api          # specific service
./tools/diagnose.sh logs api 500      # last 500 lines
./tools/diagnose.sh logs worker 200   # worker logs

Container Status

./tools/diagnose.sh ps

System Resources

./tools/diagnose.sh disk    # disk usage breakdown
./tools/diagnose.sh mem     # memory usage
./tools/diagnose.sh top     # top processes by memory

File Operations

./tools/diagnose.sh ls <path>              # list directory
./tools/diagnose.sh cat <file>             # view file
./tools/diagnose.sh tail <file> [lines]    # tail file (default: 50)
./tools/diagnose.sh grep <pattern> <path>  # search files

Database Queries

Run read-only SQL queries (SELECT only):

./tools/diagnose.sh db "SELECT count(*) FROM source_item"
./tools/diagnose.sh db "SELECT * FROM source_item ORDER BY created_at DESC LIMIT 5"
./tools/diagnose.sh db "SELECT status, count(*) FROM source_item GROUP BY status"

Non-SELECT queries are blocked for safety.

Common Diagnostics

Check why API is failing:

./tools/diagnose.sh logs api 200

Check worker processing:

./tools/diagnose.sh logs worker 100

Database stats:

./tools/diagnose.sh db "SELECT count(*) FROM source_item"
./tools/diagnose.sh db "SELECT modality, count(*) FROM source_item GROUP BY modality"

Find recent errors:

./tools/diagnose.sh status   # includes error grep
./tools/diagnose.sh grep "ERROR" src/

Check disk space:

./tools/diagnose.sh disk