Yes, New Relic AI monitoring can exclude sensitive data when monitoring DeepSeek-R1 applications. This capability is crucial for maintaining data privacy and compliance while leveraging AI models like DeepSeek-R1. Here's how it works:
1. Drop Filters: New Relic provides a feature called drop filters, which allow users to selectively exclude specific data types from monitoring. This is particularly useful for excluding personally identifiable information (PII) or other sensitive data. To set up drop filters, users navigate to the AI Monitoring section within the New Relic dashboard, select Drop Filters, and create filters using New Relic Query Language (NRQL) queries. These filters can target specific attributes within the six events offered by AI monitoring[1][9].
2. Regex for Data Exclusion: Users can utilize regular expressions (regex) to match and exclude sensitive information. For example, regex patterns can be used to identify and drop data such as credit card numbers or Social Security numbers. This ensures that only non-sensitive data is sent to New Relic for analysis[9].
3. Agent Configuration: Another option is to disable the recording of certain event data by configuring the agent settings. For instance, disabling `ai_monitoring.record_content.enabled` prevents the transmission of end-user prompts and AI responses to New Relic. This approach provides a complete opt-out option for sensitive data transmission[9].
4. Security and Privacy: New Relic emphasizes robust security features to ensure compliance and user privacy. The platform is designed to handle sensitive data securely, and users have full control over what data is monitored and transmitted[1][4].
By integrating these features, New Relic AI monitoring supports the secure and efficient deployment of DeepSeek models, allowing businesses to leverage AI while protecting sensitive information. This integration is part of New Relic's broader strategy to simplify the development, deployment, and monitoring of generative AI applications, making AI more accessible and cost-effective for enterprises[2][10].
Citations:
[1] https://newrelic.com/blog/nerdlog/ai-monitoring-ga
[2] https://newrelic.com/press-release/20250203
[3] https://docs.newrelic.com/docs/logs/ui-data/obfuscation-ui/
[4] https://newrelic.com/blog/how-to-relic/deploy-deepseek-models-locally-and-monitor-with-new-relic-ai-monitoring
[5] https://radicaldatascience.wordpress.com/tag/deepseek/
[6] https://www.reddit.com/r/SteamDeck/comments/12jk62l/i_have_managed_to_compile_and_launch_the_offline/
[7] https://newrelic.com/blog/how-to-relic/protect-your-log-data-from-security-breaches
[8] https://www.linkedin.com/posts/gariano_deploy-deepseek-models-locally-and-monitor-activity-7292645533943226369-BtJH
[9] https://docs.newrelic.com/docs/ai-monitoring/drop-sensitive-data/
[10] https://www.computerweekly.com/news/366618774/New-Relic-extends-observability-to-DeepSeek
[11] https://docs.newrelic.com/docs/security/security-privacy/data-privacy/data-privacy-new-relic/