Skip to main content

Database Schema

Rug Radar uses Postgres for watchlists, snapshots, alerts, and tool-event logging.

The schema is defined in migrations/001_init.sql.

Tables​

watchlists​

Stores one active rule set per user_id and token_address.

Key fields:

  • user_id
  • token_address
  • rules_json
  • active
  • timestamps

Important behavior:

  • watch_token upserts on (user_id, token_address)

token_snapshots​

Stores token analysis snapshots over time.

Key fields:

  • token_address
  • snapshot_json
  • risk_score
  • liquidity
  • top10_pct
  • fetched_at

Important behavior:

  • get_token_changes loads the latest snapshot and then inserts a new one

alerts​

Stores triggered alert records tied to a watchlist.

Key fields:

  • watchlist_id
  • token_address
  • trigger_type
  • trigger_detail
  • dedupe_key
  • sent_at

Important behavior:

  • dedupe_key is unique
  • Redis is used to reduce duplicate alert insertion attempts within the current dedupe window

tool_events​

Stores tool usage events for observability and future analytics.

Key fields:

  • tool_name
  • user_id
  • token_address
  • meta_json
  • created_at

What Is Not Stored​

  • continuous market ticks
  • full wallet histories outside current tool calls
  • user auth records
  • push-delivery state for alerts

Operational Implications​

  • if Postgres is unavailable, watch and change-tracking features will fail
  • token and wallet analysis without persistence can still work