Data Watchdog
Automatically detect and recover from stale or stagnant market data feeds
Overview
The Data Watchdog monitors a reference instrument's price feed (usually the ES front-month contract) for each broker connection and triggers a reconnect when data stops moving. This catches a class of problems that a simple connection status check misses. Your connection may show as "Connected" in NinjaTrader, but the price data has silently stopped updating.
Without the Watchdog, you could be running automated strategies against frozen quotes without realizing it.
Data Watchdog is available in XT Add-On v1.12.0+. It is configured per connection through the XT Web Dashboard or the API.
How It Works
The Watchdog subscribes to market data for a reference instrument (default: front-month ES) and tracks the timestamp of the last price tick for each monitored connection. Every few seconds, it compares the elapsed time since the last tick against your configured timeout.
If the data has been stagnant longer than the timeout and the market is currently open, the Watchdog triggers a recovery. If Auto-Reconnect is enabled for that connection, the Watchdog delegates the recovery to it. If Auto-Reconnect is not enabled, the Watchdog handles the disconnect and reconnect directly with its own attempt limits.
The Watchdog automatically pauses checking when the market is closed. Weekends, holidays, and outside-of-session hours do not trigger false alarms.
Configuration
Each connection can have its own Watchdog settings:
Connection
—
The broker connection to monitor.
Instrument
ES (front month)
The reference instrument to watch for price updates. It is not required for users to set this value.
Timeout (seconds)
60
How long the data must be stagnant before triggering a recovery.
Interaction with Auto-Reconnect
When both features are enabled for the same connection, they coordinate:
The Data Watchdog detects the problem (stale data) and initiates a disconnect.
Auto-Reconnect handles the actual reconnection with its delay, retry budget, and deduplication guards.
The Watchdog does not maintain its own separate attempt counter in this case — it defers entirely to Auto-Reconnect's budget to avoid double-counting.
If Auto-Reconnect is not enabled, the Watchdog runs its own recovery loop with the same maximum attempt limits.
Last updated