llm-sentinel-alerting

Crates.iollm-sentinel-alerting
lib.rsllm-sentinel-alerting
version0.1.0
created_at2025-11-06 08:11:23.921156+00
updated_at2025-11-06 08:11:23.921156+00
descriptionRabbitMQ and webhook alerting with deduplication for LLM-Sentinel anomaly detection
homepagehttps://github.com/globalbusinessadvisors/llm-sentinel
repositoryhttps://github.com/globalbusinessadvisors/llm-sentinel
max_upload_size
id1919280
size140,076
GBA (globalbusinessadvisors)

documentation

README

llm-sentinel-alerting

RabbitMQ and webhook alerting with deduplication for LLM-Sentinel.

Overview

Flexible alerting system with multiple delivery channels:

  • RabbitMQ: Topic-based routing with severity levels
  • Webhooks: HTTP POST with HMAC-SHA256 signatures
  • Deduplication: 5-minute window to prevent alert storms
  • Retry Logic: Exponential backoff for reliable delivery

Features

  • Multiple alert channels (RabbitMQ, webhooks)
  • Automatic deduplication
  • HMAC signature verification
  • Persistent message delivery
  • Configurable retry policies
  • Topic-based routing by severity

Usage

[dependencies]
llm-llm-sentinel-alerting = "0.1.0"

Example

use llm_sentinel_alerting::{AlertingManager, AlertingConfig};

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let config = AlertingConfig {
        rabbitmq_url: "amqp://localhost:5672".to_string(),
        exchange: "sentinel.alerts".to_string(),
        ..Default::default()
    };

    let manager = AlertingManager::new(config).await?;
    manager.send_alert(&anomaly).await?;

    Ok(())
}

License

Apache-2.0

Commit count: 0

cargo fmt