apiVersion: v1
data:
  api-gateway.json: '{"title":"API Gateway","uid":"api-gateway","schemaVersion":39,"timezone":"browser","time":{"from":"now-6h","to":"now"},"refresh":"30s","tags":["api","gateway","llm"],"panels":[{"id":1,"title":"Gateway
    Health","type":"row","collapsed":false,"gridPos":{"h":1,"w":24,"x":0,"y":0},"panels":[{"id":2,"title":"Gateway
    Pods Ready","type":"stat","gridPos":{"h":4,"w":4,"x":0,"y":1},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"short","thresholds":{"mode":"absolute","steps":[{"value":null,"color":"red"},{"value":3,"color":"green"}]},"color":{"mode":"thresholds"}}},"targets":[{"expr":"sum(kube_pod_status_ready{namespace=\"api\",condition=\"true\"})"}]},{"id":3,"title":"Probe:
    healthz","type":"stat","gridPos":{"h":4,"w":4,"x":4,"y":1},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"short","mappings":[{"type":"value","options":{"0":{"text":"DOWN","color":"red"},"1":{"text":"UP","color":"green"}}}]}},"targets":[{"expr":"probe_success{instance=~\".*api.riotpiao.com/healthz\"}"}]},{"id":4,"title":"Probe
    Latency","type":"timeseries","gridPos":{"h":8,"w":8,"x":8,"y":1},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"s"}},"targets":[{"expr":"probe_duration_seconds{instance=~\".*api.riotpiao.com.*\"}","legendFormat":"{{instance}}"}]}]},{"id":10,"title":"Ingress
    Traffic (nginx)","type":"row","collapsed":true,"gridPos":{"h":1,"w":24,"x":0,"y":1},"panels":[{"id":11,"title":"Request
    Rate by Status","type":"timeseries","gridPos":{"h":8,"w":8,"x":0,"y":2},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"short"}},"targets":[{"expr":"sum(rate(nginx_ingress_controller_requests{ingress=\"api\"}[5m]))
    by (status)","legendFormat":"{{status}}"}]},{"id":12,"title":"Error Rate %","type":"timeseries","gridPos":{"h":8,"w":8,"x":8,"y":2},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"percent"}},"targets":[{"expr":"sum(rate(nginx_ingress_controller_requests{ingress=\"api\",status=~\"5..\"}[5m]))
    / sum(rate(nginx_ingress_controller_requests{ingress=\"api\"}[5m])) * 100","legendFormat":"5xx"},{"expr":"sum(rate(nginx_ingress_controller_requests{ingress=\"api\",status=~\"4..\"}[5m]))
    / sum(rate(nginx_ingress_controller_requests{ingress=\"api\"}[5m])) * 100","legendFormat":"4xx"}]},{"id":13,"title":"Latency
    p50/p95/p99","type":"timeseries","gridPos":{"h":8,"w":8,"x":16,"y":2},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"s"}},"targets":[{"expr":"histogram_quantile(0.50,
    sum(rate(nginx_ingress_controller_request_duration_seconds_bucket{ingress=\"api\"}[5m]))
    by (le))","legendFormat":"p50"},{"expr":"histogram_quantile(0.95, sum(rate(nginx_ingress_controller_request_duration_seconds_bucket{ingress=\"api\"}[5m]))
    by (le))","legendFormat":"p95"},{"expr":"histogram_quantile(0.99, sum(rate(nginx_ingress_controller_request_duration_seconds_bucket{ingress=\"api\"}[5m]))
    by (le))","legendFormat":"p99"}]}]},{"id":20,"title":"LLM Serving","type":"row","collapsed":true,"gridPos":{"h":1,"w":24,"x":0,"y":2},"panels":[{"id":21,"title":"LLM
    Pods Ready","type":"stat","gridPos":{"h":4,"w":4,"x":0,"y":3},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"short"}},"targets":[{"expr":"sum(kube_pod_status_ready{namespace=\"llm-serving\",condition=\"true\"})"}]},{"id":22,"title":"CPU
    by Predictor","type":"timeseries","gridPos":{"h":8,"w":8,"x":4,"y":3},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"short"}},"targets":[{"expr":"sum(rate(container_cpu_usage_seconds_total{namespace=\"llm-serving\"}[5m]))
    by (pod)","legendFormat":"{{pod}}"}]},{"id":23,"title":"Memory by Predictor","type":"timeseries","gridPos":{"h":8,"w":8,"x":12,"y":3},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"bytes"}},"targets":[{"expr":"sum(container_memory_working_set_bytes{namespace=\"llm-serving\"})
    by (pod)","legendFormat":"{{pod}}"}]},{"id":24,"title":"Predictor Restarts","type":"timeseries","gridPos":{"h":8,"w":4,"x":20,"y":3},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"short"}},"targets":[{"expr":"sum(rate(kube_pod_container_status_restarts_total{namespace=\"llm-serving\"}[15m]))
    by (pod)","legendFormat":"{{pod}}"}]}]},{"id":30,"title":"Gateway Resources","type":"row","collapsed":true,"gridPos":{"h":1,"w":24,"x":0,"y":3},"panels":[{"id":31,"title":"CPU
    by Gateway Pod","type":"timeseries","gridPos":{"h":8,"w":8,"x":0,"y":4},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"short"}},"targets":[{"expr":"sum(rate(container_cpu_usage_seconds_total{namespace=\"api\"}[5m]))
    by (pod)","legendFormat":"{{pod}}"}]},{"id":32,"title":"Memory by Gateway Pod","type":"timeseries","gridPos":{"h":8,"w":8,"x":8,"y":4},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"bytes"}},"targets":[{"expr":"sum(container_memory_working_set_bytes{namespace=\"api\"})
    by (pod)","legendFormat":"{{pod}}"}]},{"id":33,"title":"Gateway Restarts","type":"timeseries","gridPos":{"h":8,"w":8,"x":16,"y":4},"datasource":{"type":"prometheus","uid":"prometheus"},"fieldConfig":{"defaults":{"unit":"short"}},"targets":[{"expr":"sum(rate(kube_pod_container_status_restarts_total{namespace=\"api\"}[15m]))
    by (pod)","legendFormat":"{{pod}}"}]}]},{"id":40,"title":"Logs","type":"row","collapsed":true,"gridPos":{"h":1,"w":24,"x":0,"y":4},"panels":[{"id":41,"title":"Gateway
    Logs","type":"logs","gridPos":{"h":10,"w":24,"x":0,"y":5},"datasource":{"type":"loki","uid":"loki"},"targets":[{"expr":"{namespace=\"api\",container=\"gateway\"}"}]},{"id":42,"title":"LLM
    Serving Logs","type":"logs","gridPos":{"h":10,"w":24,"x":0,"y":15},"datasource":{"type":"loki","uid":"loki"},"targets":[{"expr":"{namespace=\"llm-serving\"}"}]}]}]}'
kind: ConfigMap
metadata:
  annotations:
    grafana_folder: API
  labels:
    grafana_dashboard: '1'
  name: api-gateway-dashboard
  namespace: logging