Pragma Monitoring

Prometheus Exporter

This service runs a prometheus exporter that exposes metrics on the /metrics route. It powers our internal grafana dashboards and alerts.

This service polls at a regular interval the database that is filled by our indexers. It then processes the data and computes the following metrics:

• time_since_last_update_seconds{network, publisher, typr}: Time since a publisher has published any data. (in seconds)
• pair_price{network, pair, source, type}: Latest price of an asset for a given source and pair. (normalized to asset's decimals)
• time_since_last_update_pair_id{network, pair, type}: Time since an update has been published for a given pair. (in seconds)
• price_deviation{network, pair, source, type}: Deviation of the price from a reference price (DefiLlama API) given source and pair. (in percents)
• price_deviation_source{network, pair, source, type}: Deviation of the price from the on-chain aggregated median price given source and pair. (in percents)
• long_tail_asset_threshold{pair, type}: Deviation threshold configuration for long tail assets. Type can be either "low" for when the pair has less than 7 sources else "high".
• long_tail_asset_source_deviation{network, pair, type}: Deviation of a source from the on-chain aggregated median price given source and pair. (in percents)
• long_tail_asset_total_sources{network, pair, type}: Current number of sources available for a given pair.
• publisher_balance{network, publisher}: Balance of a publisher. (in ETH)

Metrics specifics to Starknet (mainnet/sepolia):

• vrf_balance{network}: Balance of the VRF contract. (in ETH)
• vrf_requests_count{network, status}: Number of VRF requests handled for a given network.
• vrf_time_since_last_handle_request{network}: Time since the last VRF request was handled for a given network.
• vrf_time_since_oldest_request_in_pending_status{network}: Time in seconds that the oldest pending VRF request has been in the pending status for a given network.

Shared Public Access

Monitoring is not publicicly available yet but databases will soon be in read-only mode.

Self-Hosting

We have created a docker-compose.yml file to help with self-hosting setup:

docker compose up -d

You can then access prometheus dashboard at http://localhost:9000 and grafana at http://localhost:3000.

Make sure to first fill the envirronement file with your own config parameters:

# The database URL the application will use to connect to the database.
DATABASE_URL='postgres://postgres:postgres@localhost:5432/postgres'

# (Optional) Defillama API Key
DEFILLAMA_API_KEY=

# Pragma API key
PRAGMA_API_KEY=

# RPC URL
RPC_URL=

# Indexer Service URL
INDEXER_SERVICE_URL=

# Config
NETWORK=testnet
ORACLE_ADDRESS=0x
VRF_ADDRESS=0x
PAIRS=BTC/USD,ETH/USD
IGNORE_SOURCES=BITSTAMP,DEFILLAMA
IGNORE_PUBLISHERS=BINANCE

# Prometheus
TELEGRAM_TOKEN=
OPSGENIE_API_KEY=

In order for the full flow to work you will need to have tables following the table schemas defined here.

You can use our indexer service on this repository to spin off your indexer in a few commands very easily.