v0.5.0

Added

• Prometheus metrics endpoint on port 5001 (/metrics) with optional HTTP Basic Auth (instancer.metrics.username / instancer.metrics.password in config)
• Deployment count gauge (instancer_deployments) backed by the database, grouped by status / category / challenge / team
• Deploy and terminate operation counters (instancer_deploy_ops_total, instancer_terminate_ops_total) with result label
• Deploy and terminate duration histograms (instancer_deploy_duration_seconds, instancer_terminate_duration_seconds) with category / challenge / team labels
• Deploy conflict counter (instancer_deploy_conflict_total) for 409 responses
• Unauthorized deploy request counter (instancer_unauthorized_deploy_requests_total) per team
• Extension operation counter (instancer_extend_ops_total) and rejection counter (instancer_extend_rejected_total) with reason label (window_not_reached, no_extensions_left, already_expired)
• Deployment lifetime histogram (instancer_deployment_lifetime_seconds) recorded at termination time
• Worker job retry counter (instancer_job_retries_total) and permanent failure counter (instancer_job_permanent_failures_total) with job_type label
• Redis queue depth gauge (instancer_queue_depth) and queue wait time histogram (instancer_job_queue_wait_seconds) — only registered when Redis is configured
• Challenge index size gauge (instancer_challenges_indexed) per category, updated on startup and on challenge reload
• Grafana monitoring stack (docker-compose.monitoring.yml) with Prometheus + Grafana, pre-provisioned with a full dashboard covering all metrics
• make monitoring-up / make monitoring-down targets

Fixed

• TCP playbook (and any playbook using {{ env | default({}) }}): challenges without an env key in deploy_parameters no longer fail with "environment must be a mapping" — env is now normalised to an empty map before being passed to Ansible