Grafana query recipes
PromQL expressions for the dashboards you’ll want when driving rcfg-sim hard. Point
Prometheus at each instance’s /metrics endpoint first.
Session throughput
Section titled “Session throughput”Successful sessions per second:
sum(rate(rcfgsim_sessions_total{result="ok"}[1m]))Error rate
Section titled “Error rate”Percentage of sessions ending in any non-ok result over 5 minutes:
sum(rate(rcfgsim_sessions_total{result!="ok"}[5m])) / sum(rate(rcfgsim_sessions_total[5m])) * 100Command latency (p95)
Section titled “Command latency (p95)”95th-percentile per-command latency, broken out by command:
histogram_quantile(0.95, sum by (command, le) (rate(rcfgsim_command_duration_seconds_bucket[5m])))Throughput in bits per second
Section titled “Throughput in bits per second”sum(rate(rcfgsim_bytes_sent_total[1m])) * 8Concurrency vs the cap
Section titled “Concurrency vs the cap”Active sessions — compare against your --max-concurrent-sessions:
sum(rcfgsim_active_sessions)Auth failure rate
Section titled “Auth failure rate”sum(rate(rcfgsim_auth_attempts_total{result="fail"}[5m])) / sum(rate(rcfgsim_auth_attempts_total[5m])) * 100Faults firing by type
Section titled “Faults firing by type”Confirm injected faults track your configured rate:
sum by (type) (rate(rcfgsim_faults_injected_total[5m]))Handshake latency (p99)
Section titled “Handshake latency (p99)”histogram_quantile(0.99, sum by (le) (rate(rcfgsim_handshake_duration_seconds_bucket[5m])))See the metrics reference for the full metric and label inventory.