Skip to content

Monitoring and tooling ​

A keeper bot is only as good as its observation. This page covers what to monitor and what tooling is available.

What to monitor ​

  • All positions with H < 150%. These are within range of becoming liquidatable on a small adverse move.
  • Pending oracle updates. A new oracle reading can shift many positions' H simultaneously.
  • Pool utilization. High utilization means high stress; opportunities appear faster.
  • Your own H. Don't liquidate yourself out of solvency.
  • Gas prices. A liquidation profitable at 30 gwei may not be profitable at 300 gwei.

How to monitor ​

banq-cli covers the basics out of the box: health-of queries an account's health factor, rates-of reports the current supply/borrow rates (and historical trajectories), and retwap --watch streams TWAP-refresh events live. See CLI and tools for the full command set.

For sub-second latency (relevant in high-volatility regimes), you'd run your own indexer that listens to the on-chain events and reads healthOf periodically — banq-cli is a reference, not a production keeper.

Position prioritisation ​

Sort underwater positions by:

  • Expected profit = bonus − gas cost.
  • Time pressure = how fast H is moving.
  • Liquidity = whether the absorbed collateral is locked or unlocked.

Sophisticated keepers run an optimisation: they pick the best slice exponent given competing keepers, lock structure, and gas economics.

Open-source tools ​

  • banq-cli — the official Deno CLI (dry-run by default). Covers liquidation, oracle refresh, health/rate queries, and ACMA inspection. See CLI and tools.

banq-cli is a starting point. Production keepers typically add proprietary monitoring, private mempool integration, and custom risk management.

What sophistication actually buys you ​

The whitepaper acknowledges that liquidation revenue concentrates at the top of the distribution: top-5 keepers in established protocols capture >80% of value. Banq's debt-assumption mechanism widens the pool of possible keepers, but not the distribution of captured value.

If you want to compete with sophisticated keepers, expect to invest in:

  • Sub-block latency monitoring.
  • Private mempool integration.
  • Lock-state-aware accounting (locked positions are worth less than NAV).
  • MEV protection.

If you want to participate at the margin without competing for the top tier, the reference implementation is enough.

Where to go next ​