Crates.io | ai-vitals |
lib.rs | ai-vitals |
version | 0.8.1 |
created_at | 2025-08-04 16:35:40.228783+00 |
updated_at | 2025-09-25 09:21:57.012385+00 |
description | A tool to monitor LLM endpoints and report status to Cronitor. |
homepage | https://doubleword.ai |
repository | https://github.com/doublewordai/model-monitor |
max_upload_size | |
id | 1780924 |
size | 133,224 |
AI Vitals is a tool designed to monitor LLM (Large Language Model) endpoints and report their status to Cronitor, a service for monitoring and alerting on the health of applications. It actively monitors OpenAI-compatible API endpoints, ensuring they are functioning correctly and reporting any issues to Cronitor for further triage.
You can use the library interface or the command-line interface (CLI) to monitor your endpoints. The CLI is particularly useful for quick checks or running in automated scripts. We have an example helm chart that uses the CLI to monitor endpoints with kubernetes CronJobs.
use ai_vitals::{Config, Monitor};
use anyhow::Result;
#[tokio::main]
async fn main() -> Result<()> {
let config = Config::parse(); // Load configuration from environment or command line
let monitor = Monitor::new(config)?;
let exit_code = monitor.run().await;
std::process::exit(exit_code);
}
You can also run the CLI directly:
cargo run --bin ai-vitals -- --cronitor-base-url "https://cronitor.link/p/your-key/your-group" \
--server-url "http://my-service" \
--model-name "embed" \
--endpoint-type "embedding" \
--monitor-name "my-embedding-model"