How does SentryDock decide what to alert me about?
Each monitor has a topic, a set of sources and a schedule. When a source produces new material, SentryDock extracts the available content and compares it with the instructions saved on that monitor.
Items that appear relevant can be summarized, translated and analysed. Duplicate or substantially similar items may be grouped so the same event does not create a stream of identical alerts. The exact analysis available depends on the monitor and plan.
What affects relevance
- The subject, entities, locations and events named in your monitoring instruction.
- Any exclusions or materiality criteria you added.
- The content that was available from the original source at check time.
- The relevance threshold and analysis settings used by the monitor.
How to reduce noisy alerts
Make the instruction more specific about what should trigger an alert and what should not. Add trusted, narrowly focused sources and remove broad sources that repeatedly produce irrelevant results. If the topic is not time-sensitive, use an hourly or daily schedule.
AI filtering can make mistakes. Open the original item before acting on a material alert, and use repeated false positives as evidence that the monitor instruction or source list needs adjustment.
If an expected alert is missing
- Confirm that the original source was public and reachable.
- Check that the event clearly matched the saved monitoring instruction.
- Review the monitor schedule and the enabled delivery destination.
- Contact support with the monitor name and original URL if the source should have matched.