Description
Expose Prometheus-compatible metrics for monitoring Sirene in production.
Proposal
/metrics endpoint with Prometheus format- Metrics: generation count, generation duration histogram, active model count, GPU memory usage, queue depth, error rate
- Grafana dashboard template included in docs
- Health check endpoint enhancement with detailed component status
Motivation
For self-hosted production deployments, observability is critical. Prometheus metrics enable alerting on errors, capacity planning based on GPU usage, and performance monitoring.
Description
Expose Prometheus-compatible metrics for monitoring Sirene in production.
Proposal
/metricsendpoint with Prometheus formatMotivation
For self-hosted production deployments, observability is critical. Prometheus metrics enable alerting on errors, capacity planning based on GPU usage, and performance monitoring.