Observe pyth data on-chain and apply some basic sanity checking on the published data.
- Python 3.7
- https://pypi.org/project/pythclient/ library at pyth-client-py.
Create a virtual environment and install all dependencies:
pushd $(pwd)
python3 -m venv ve
. ve/bin/activate
Install the pyth-observer dependencies:
pip install -r requirements.txtRun the unit tests by running pytest from the repository root.
| Name | Default | Description |
|---|---|---|
PYTH_OBSERVER_SLACK_WEBHOOK_URL |
None | URL of the Slack incoming webhook for notifications |
PYTH_OBSERVER_PRICE_DEVIATION_THRESHOLD |
6 | Percentage between the published price and aggregate |
PYTH_OBSERVER_TWAP_VS_AGGREGATE_THRESHOLD |
10 | Max TWAP and Aggregate prices should be apart |
PYTH_OBSERVER_STOP_PUBLISHING_MIN_SLOTS |
600 | Min slots before stop-publishing alert fires |
PYTH_OBSERVER_STOP_PUBLISHING_MAX_SLOTS |
1000 | Max slots behind a stop-publishing alert fires on |
PYTH_OBSERVER_PRICE_DEVIATION_COINGECKO |
5 | Percentage between aggregate price and CoinGecko price |
By default, pyth observer only logs to stderr:
./observer.py -l debug --network=mainnetUse the trace log level to see exactly what is going on.
To send alerts to slack, you need to create an app and get the incoming webhook url. Export it as the variable PYTH_OBSERVER_SLACK_WEBHOOK_URL:
export PYTH_OBSERVER_SLACK_WEBHOOK_URL=https://hooks.slack.com/services/T0GPR2P4K/B02J164R5MF/XYZ123LMAOZOMGBBQWTF