refactor(server): prune otel dashboard and active http request
This commit is contained in:
RainbowBird
@@ -1,279 +0,0 @@
|
||||
{
|
||||
"annotations": { "list": [] },
|
||||
"editable": true,
|
||||
"fiscalYearStartMonth": 0,
|
||||
"graphTooltip": 1,
|
||||
"links": [],
|
||||
"panels": [
|
||||
{
|
||||
"collapsed": false,
|
||||
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 0 },
|
||||
"id": 100,
|
||||
"title": "Chat",
|
||||
"type": "row"
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 50 },
|
||||
{ "color": "red", "value": 200 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 0, "y": 1 },
|
||||
"id": 1,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Chat Syncs (1h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(chat_sync_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[1h]))",
|
||||
"legendFormat": "syncs"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 100 },
|
||||
{ "color": "red", "value": 500 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 8, "y": 1 },
|
||||
"id": 2,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Messages (1h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(chat_messages_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[1h]))",
|
||||
"legendFormat": "messages"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20,
|
||||
"stacking": { "mode": "none" }
|
||||
},
|
||||
"unit": "reqps"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 0, "y": 9 },
|
||||
"id": 3,
|
||||
"title": "Chat Sync Rate by Type",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum by (chat_type) (rate(chat_sync_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "{{chat_type}}"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20,
|
||||
"stacking": { "mode": "none" }
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 12, "y": 9 },
|
||||
"id": 4,
|
||||
"title": "Message Volume Over Time",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(rate(chat_messages_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "messages/s"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"collapsed": false,
|
||||
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 17 },
|
||||
"id": 101,
|
||||
"title": "Characters",
|
||||
"type": "row"
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 10 },
|
||||
{ "color": "red", "value": 50 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 0, "y": 18 },
|
||||
"id": 5,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Characters Created (24h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(character_created_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[24h]))",
|
||||
"legendFormat": "created"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 5 },
|
||||
{ "color": "red", "value": 20 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 8, "y": 18 },
|
||||
"id": 6,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Characters Deleted (24h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(character_deleted_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[24h]))",
|
||||
"legendFormat": "deleted"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "bars",
|
||||
"fillOpacity": 80,
|
||||
"stacking": { "mode": "normal" }
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 0, "y": 26 },
|
||||
"id": 7,
|
||||
"title": "Engagement Actions",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum by (action) (rate(character_engagement_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "{{action}}"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20,
|
||||
"stacking": { "mode": "none" }
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 12, "y": 26 },
|
||||
"id": 8,
|
||||
"title": "Character Creation Over Time",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(rate(character_created_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "created/s"
|
||||
}
|
||||
]
|
||||
}
|
||||
],
|
||||
"schemaVersion": 39,
|
||||
"tags": ["airi", "engagement", "grafana-cloud"],
|
||||
"templating": {
|
||||
"list": [
|
||||
{
|
||||
"current": { "selected": false, "text": "production", "value": "production" },
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"definition": "label_values(http_server_request_duration_milliseconds_count, deployment_environment)",
|
||||
"hide": 0,
|
||||
"includeAll": true,
|
||||
"multi": false,
|
||||
"name": "env",
|
||||
"query": "label_values(http_server_request_duration_milliseconds_count, deployment_environment)",
|
||||
"refresh": 2,
|
||||
"type": "query"
|
||||
},
|
||||
{
|
||||
"current": { "selected": false, "text": "server", "value": "server" },
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"definition": "label_values(http_server_request_duration_milliseconds_count{deployment_environment=~\"$env\"}, service_name)",
|
||||
"hide": 0,
|
||||
"includeAll": true,
|
||||
"multi": true,
|
||||
"name": "service",
|
||||
"query": "label_values(http_server_request_duration_milliseconds_count{deployment_environment=~\"$env\"}, service_name)",
|
||||
"refresh": 2,
|
||||
"type": "query"
|
||||
}
|
||||
]
|
||||
},
|
||||
"time": { "from": "now-24h", "to": "now" },
|
||||
"title": "AIRI Engagement (Cloud)",
|
||||
"uid": "airi-engagement-cloud"
|
||||
}
|
||||
@@ -1,319 +0,0 @@
|
||||
{
|
||||
"annotations": { "list": [] },
|
||||
"editable": true,
|
||||
"fiscalYearStartMonth": 0,
|
||||
"graphTooltip": 1,
|
||||
"links": [],
|
||||
"panels": [
|
||||
{
|
||||
"collapsed": false,
|
||||
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 0 },
|
||||
"id": 100,
|
||||
"title": "Stripe",
|
||||
"type": "row"
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 10 },
|
||||
{ "color": "red", "value": 50 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 6, "x": 0, "y": 1 },
|
||||
"id": 1,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Checkouts Created (24h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(stripe_checkout_created_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[24h]))",
|
||||
"legendFormat": "created"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 5 },
|
||||
{ "color": "red", "value": 30 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 6, "x": 6, "y": 1 },
|
||||
"id": 2,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Checkouts Completed (24h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(stripe_checkout_completed_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[24h]))",
|
||||
"legendFormat": "completed"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "red", "value": null },
|
||||
{ "color": "yellow", "value": 1 },
|
||||
{ "color": "red", "value": 5 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 6, "x": 12, "y": 1 },
|
||||
"id": 3,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Payment Failures (24h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(stripe_payment_failed_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[24h]))",
|
||||
"legendFormat": "failures"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "red", "value": null },
|
||||
{ "color": "yellow", "value": 0.5 },
|
||||
{ "color": "green", "value": 0.8 }
|
||||
]
|
||||
},
|
||||
"unit": "percentunit"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 6, "x": 18, "y": 1 },
|
||||
"id": 4,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Conversion Rate",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(stripe_checkout_completed_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[24h])) / sum(increase(stripe_checkout_created_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[24h]))",
|
||||
"legendFormat": "conversion"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20,
|
||||
"stacking": { "mode": "none" }
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 0, "y": 9 },
|
||||
"id": 5,
|
||||
"title": "Checkout Funnel Over Time",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(rate(stripe_checkout_created_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "created/s"
|
||||
},
|
||||
{
|
||||
"expr": "sum(rate(stripe_checkout_completed_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "completed/s"
|
||||
},
|
||||
{
|
||||
"expr": "sum(rate(stripe_payment_failed_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "failed/s"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "bars",
|
||||
"fillOpacity": 80,
|
||||
"stacking": { "mode": "normal" }
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 12, "y": 9 },
|
||||
"id": 6,
|
||||
"title": "Subscription Events",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum by (event_type) (rate(stripe_subscription_event_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "{{event_type}}"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20,
|
||||
"stacking": { "mode": "none" }
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 24, "x": 0, "y": 17 },
|
||||
"id": 7,
|
||||
"title": "Stripe Events by Type",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum by (event_type) (rate(stripe_events_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "{{event_type}}"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"collapsed": false,
|
||||
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 25 },
|
||||
"id": 101,
|
||||
"title": "Flux",
|
||||
"type": "row"
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 5 },
|
||||
{ "color": "red", "value": 20 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 0, "y": 26 },
|
||||
"id": 8,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Insufficient Balance (1h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(flux_insufficient_balance_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[1h]))",
|
||||
"legendFormat": "insufficient"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20,
|
||||
"stacking": { "mode": "none" }
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 16, "x": 8, "y": 26 },
|
||||
"id": 9,
|
||||
"title": "Insufficient Balance Over Time",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(rate(flux_insufficient_balance_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "insufficient/s"
|
||||
}
|
||||
]
|
||||
}
|
||||
],
|
||||
"schemaVersion": 39,
|
||||
"tags": ["airi", "revenue", "stripe", "grafana-cloud"],
|
||||
"templating": {
|
||||
"list": [
|
||||
{
|
||||
"current": { "selected": false, "text": "production", "value": "production" },
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"definition": "label_values(http_server_request_duration_milliseconds_count, deployment_environment)",
|
||||
"hide": 0,
|
||||
"includeAll": true,
|
||||
"multi": false,
|
||||
"name": "env",
|
||||
"query": "label_values(http_server_request_duration_milliseconds_count, deployment_environment)",
|
||||
"refresh": 2,
|
||||
"type": "query"
|
||||
},
|
||||
{
|
||||
"current": { "selected": false, "text": "server", "value": "server" },
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"definition": "label_values(http_server_request_duration_milliseconds_count{deployment_environment=~\"$env\"}, service_name)",
|
||||
"hide": 0,
|
||||
"includeAll": true,
|
||||
"multi": true,
|
||||
"name": "service",
|
||||
"query": "label_values(http_server_request_duration_milliseconds_count{deployment_environment=~\"$env\"}, service_name)",
|
||||
"refresh": 2,
|
||||
"type": "query"
|
||||
}
|
||||
]
|
||||
},
|
||||
"time": { "from": "now-7d", "to": "now" },
|
||||
"title": "AIRI Revenue (Cloud)",
|
||||
"uid": "airi-revenue-cloud"
|
||||
}
|
||||
File diff suppressed because it is too large
Load Diff
@@ -1,503 +0,0 @@
|
||||
{
|
||||
"annotations": {
|
||||
"list": []
|
||||
},
|
||||
"editable": true,
|
||||
"fiscalYearStartMonth": 0,
|
||||
"graphTooltip": 1,
|
||||
"links": [],
|
||||
"panels": [
|
||||
{
|
||||
"collapsed": false,
|
||||
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 0 },
|
||||
"id": 100,
|
||||
"title": "HTTP Overview",
|
||||
"type": "row"
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 100 },
|
||||
{ "color": "red", "value": 500 }
|
||||
]
|
||||
},
|
||||
"unit": "reqps"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 0, "y": 1 },
|
||||
"id": 1,
|
||||
"options": {
|
||||
"reduceOptions": {
|
||||
"calcs": ["lastNotNull"],
|
||||
"fields": "",
|
||||
"values": false
|
||||
},
|
||||
"orientation": "auto",
|
||||
"textMode": "auto",
|
||||
"colorMode": "value",
|
||||
"graphMode": "area",
|
||||
"justifyMode": "auto"
|
||||
},
|
||||
"title": "Request Rate",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(rate(airi_http_server_request_duration_seconds_count{service_name=~\"$service\"}[$__rate_interval]))",
|
||||
"legendFormat": "req/s"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 0.01 },
|
||||
{ "color": "red", "value": 0.05 }
|
||||
]
|
||||
},
|
||||
"unit": "percentunit"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 8, "y": 1 },
|
||||
"id": 2,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Error Rate (5xx)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(rate(airi_http_server_request_duration_seconds_count{service_name=~\"$service\", http_status_code=~\"5..\"}[$__rate_interval])) / sum(rate(airi_http_server_request_duration_seconds_count{service_name=~\"$service\"}[$__rate_interval]))",
|
||||
"legendFormat": "error rate"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 0.2 },
|
||||
{ "color": "red", "value": 1.0 }
|
||||
]
|
||||
},
|
||||
"unit": "s"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 16, "y": 1 },
|
||||
"id": 3,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "P95 Latency (excl. LLM)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "histogram_quantile(0.95, sum(rate(airi_http_server_request_duration_seconds_bucket{service_name=~\"$service\", http_route!~\"/api/v1/openai/.*\"}[$__rate_interval])) by (le))",
|
||||
"legendFormat": "p95"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20,
|
||||
"stacking": { "mode": "none" }
|
||||
},
|
||||
"unit": "reqps"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 0, "y": 9 },
|
||||
"id": 4,
|
||||
"title": "Request Rate by Route",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum by (http_route) (rate(airi_http_server_request_duration_seconds_count{service_name=~\"$service\"}[$__rate_interval]))",
|
||||
"legendFormat": "{{http_route}}"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20
|
||||
},
|
||||
"unit": "s"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 12, "y": 9 },
|
||||
"id": 5,
|
||||
"title": "Latency by Route (P95)",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "histogram_quantile(0.95, sum by (le, http_route) (rate(airi_http_server_request_duration_seconds_bucket{service_name=~\"$service\"}[$__rate_interval])))",
|
||||
"legendFormat": "{{http_route}}"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "bars",
|
||||
"fillOpacity": 80,
|
||||
"stacking": { "mode": "normal" }
|
||||
},
|
||||
"unit": "reqps"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 24, "x": 0, "y": 17 },
|
||||
"id": 6,
|
||||
"title": "Response Status Codes",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum by (http_status_code) (rate(airi_http_server_request_duration_seconds_count{service_name=~\"$service\"}[$__rate_interval]))",
|
||||
"legendFormat": "{{http_status_code}}"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"collapsed": false,
|
||||
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 25 },
|
||||
"id": 101,
|
||||
"title": "Node.js Runtime",
|
||||
"type": "row"
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20
|
||||
},
|
||||
"unit": "bytes"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 0, "y": 26 },
|
||||
"id": 7,
|
||||
"title": "Heap Memory Usage",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "airi_process_runtime_nodejs_memory_heap_total{service_name=~\"$service\"}",
|
||||
"legendFormat": "heap total"
|
||||
},
|
||||
{
|
||||
"expr": "airi_process_runtime_nodejs_memory_heap_used{service_name=~\"$service\"}",
|
||||
"legendFormat": "heap used"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20
|
||||
},
|
||||
"unit": "bytes"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 12, "y": 26 },
|
||||
"id": 8,
|
||||
"title": "RSS & External Memory",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "airi_process_runtime_nodejs_memory_rss{service_name=~\"$service\"}",
|
||||
"legendFormat": "RSS"
|
||||
},
|
||||
{
|
||||
"expr": "airi_process_runtime_nodejs_memory_array_buffers{service_name=~\"$service\"}",
|
||||
"legendFormat": "array buffers"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 0, "y": 34 },
|
||||
"id": 9,
|
||||
"title": "Event Loop Lag",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "airi_process_runtime_nodejs_event_loop_lag_p99{service_name=~\"$service\"}",
|
||||
"legendFormat": "p99 lag"
|
||||
},
|
||||
{
|
||||
"expr": "airi_process_runtime_nodejs_event_loop_lag_p50{service_name=~\"$service\"}",
|
||||
"legendFormat": "p50 lag"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 12, "y": 34 },
|
||||
"id": 10,
|
||||
"title": "Active Handles & Requests",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "airi_process_runtime_nodejs_active_handles_total{service_name=~\"$service\"}",
|
||||
"legendFormat": "active handles"
|
||||
},
|
||||
{
|
||||
"expr": "airi_process_runtime_nodejs_active_requests_total{service_name=~\"$service\"}",
|
||||
"legendFormat": "active requests"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"collapsed": false,
|
||||
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 42 },
|
||||
"id": 102,
|
||||
"title": "PostgreSQL",
|
||||
"type": "row"
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20
|
||||
},
|
||||
"unit": "s"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 0, "y": 43 },
|
||||
"id": 11,
|
||||
"title": "DB Query Duration (P95)",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "histogram_quantile(0.95, sum by (le, db_operation) (rate(airi_db_client_operation_duration_seconds_bucket{service_name=~\"$service\"}[$__rate_interval])))",
|
||||
"legendFormat": "{{db_operation}}"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20
|
||||
},
|
||||
"unit": "ops"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 12, "y": 43 },
|
||||
"id": 12,
|
||||
"title": "DB Query Rate",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum by (db_operation) (rate(airi_db_client_operation_duration_seconds_count{service_name=~\"$service\"}[$__rate_interval]))",
|
||||
"legendFormat": "{{db_operation}}"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 24, "x": 0, "y": 51 },
|
||||
"id": 13,
|
||||
"title": "DB Connection Pool",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "airi_db_client_connections_usage{service_name=~\"$service\", state=\"used\"}",
|
||||
"legendFormat": "used"
|
||||
},
|
||||
{
|
||||
"expr": "airi_db_client_connections_usage{service_name=~\"$service\", state=\"idle\"}",
|
||||
"legendFormat": "idle"
|
||||
},
|
||||
{
|
||||
"expr": "airi_db_client_connections_pending_requests{service_name=~\"$service\"}",
|
||||
"legendFormat": "pending"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"collapsed": false,
|
||||
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 59 },
|
||||
"id": 103,
|
||||
"title": "Redis",
|
||||
"type": "row"
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20
|
||||
},
|
||||
"unit": "s"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 0, "y": 60 },
|
||||
"id": 14,
|
||||
"title": "Redis Command Duration (P95)",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "histogram_quantile(0.95, sum by (le, db_operation) (rate(airi_db_client_operation_duration_seconds_bucket{service_name=~\"$service\", db_system=\"redis\"}[$__rate_interval])))",
|
||||
"legendFormat": "{{db_operation}}"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20
|
||||
},
|
||||
"unit": "ops"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 12, "y": 60 },
|
||||
"id": 15,
|
||||
"title": "Redis Command Rate",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum by (db_operation) (rate(airi_db_client_operation_duration_seconds_count{service_name=~\"$service\", db_system=\"redis\"}[$__rate_interval]))",
|
||||
"legendFormat": "{{db_operation}}"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"collapsed": false,
|
||||
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 68 },
|
||||
"id": 104,
|
||||
"title": "Application Logs",
|
||||
"type": "row"
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "loki", "uid": "loki" },
|
||||
"gridPos": { "h": 10, "w": 24, "x": 0, "y": 69 },
|
||||
"id": 16,
|
||||
"title": "Application Logs",
|
||||
"type": "logs",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "{service_name=~\"$service\"} |= ``",
|
||||
"legendFormat": ""
|
||||
}
|
||||
],
|
||||
"options": {
|
||||
"showTime": true,
|
||||
"showLabels": true,
|
||||
"showCommonLabels": false,
|
||||
"wrapLogMessage": true,
|
||||
"prettifyLogMessage": false,
|
||||
"enableLogDetails": true,
|
||||
"sortOrder": "Descending"
|
||||
}
|
||||
}
|
||||
],
|
||||
"schemaVersion": 39,
|
||||
"tags": ["airi", "observability"],
|
||||
"templating": {
|
||||
"list": [
|
||||
{
|
||||
"current": { "selected": false, "text": "Prometheus", "value": "Prometheus" },
|
||||
"hide": 0,
|
||||
"includeAll": false,
|
||||
"name": "datasource",
|
||||
"options": [],
|
||||
"query": "prometheus",
|
||||
"refresh": 1,
|
||||
"type": "datasource"
|
||||
},
|
||||
{
|
||||
"current": { "selected": false, "text": "airi-server", "value": "airi-server" },
|
||||
"datasource": { "type": "prometheus", "uid": "${datasource}" },
|
||||
"definition": "label_values(airi_http_server_request_duration_seconds_count, service_name)",
|
||||
"hide": 0,
|
||||
"includeAll": true,
|
||||
"multi": true,
|
||||
"name": "service",
|
||||
"query": "label_values(airi_http_server_request_duration_seconds_count, service_name)",
|
||||
"refresh": 2,
|
||||
"type": "query"
|
||||
}
|
||||
]
|
||||
},
|
||||
"time": { "from": "now-1h", "to": "now" },
|
||||
"title": "AIRI Server Overview",
|
||||
"uid": "airi-server-overview"
|
||||
}
|
||||
@@ -1,421 +0,0 @@
|
||||
{
|
||||
"annotations": { "list": [] },
|
||||
"editable": true,
|
||||
"fiscalYearStartMonth": 0,
|
||||
"graphTooltip": 1,
|
||||
"links": [],
|
||||
"panels": [
|
||||
{
|
||||
"collapsed": false,
|
||||
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 0 },
|
||||
"id": 100,
|
||||
"title": "Registration",
|
||||
"type": "row"
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 10 },
|
||||
{ "color": "red", "value": 50 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 0, "y": 1 },
|
||||
"id": 1,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"], "fields": "", "values": false },
|
||||
"orientation": "auto",
|
||||
"textMode": "auto",
|
||||
"colorMode": "value",
|
||||
"graphMode": "area",
|
||||
"justifyMode": "auto"
|
||||
},
|
||||
"title": "New Users (24h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(user_registered_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[24h]))",
|
||||
"legendFormat": "new users"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 5 },
|
||||
{ "color": "red", "value": 20 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 8, "y": 1 },
|
||||
"id": 2,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "New Users (1h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(user_registered_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[1h]))",
|
||||
"legendFormat": "new users"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 0.1 },
|
||||
{ "color": "red", "value": 1 }
|
||||
]
|
||||
},
|
||||
"unit": "reqps"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 16, "y": 1 },
|
||||
"id": 3,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Registration Rate",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(rate(user_registered_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "reg/s"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "bars",
|
||||
"fillOpacity": 80,
|
||||
"stacking": { "mode": "none" }
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 24, "x": 0, "y": 9 },
|
||||
"id": 4,
|
||||
"title": "Registrations Over Time",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(user_registered_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "new users"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"collapsed": false,
|
||||
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 17 },
|
||||
"id": 101,
|
||||
"title": "Sessions",
|
||||
"type": "row"
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 50 },
|
||||
{ "color": "red", "value": 200 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 0, "y": 18 },
|
||||
"id": 5,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Active Sessions",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(user_active_sessions{service_name=~\"$service\", deployment_environment=~\"$env\"})",
|
||||
"legendFormat": "active"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 50 },
|
||||
{ "color": "red", "value": 200 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 8, "y": 18 },
|
||||
"id": 6,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Logins (1h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(user_login_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[1h]))",
|
||||
"legendFormat": "logins"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20,
|
||||
"stacking": { "mode": "none" }
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 0, "y": 26 },
|
||||
"id": 7,
|
||||
"title": "Login Rate Over Time",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(rate(user_login_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "logins/s"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20,
|
||||
"stacking": { "mode": "none" }
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 12, "x": 12, "y": 26 },
|
||||
"id": 8,
|
||||
"title": "Active Sessions Over Time",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(user_active_sessions{service_name=~\"$service\", deployment_environment=~\"$env\"})",
|
||||
"legendFormat": "active sessions"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"collapsed": false,
|
||||
"gridPos": { "h": 1, "w": 24, "x": 0, "y": 34 },
|
||||
"id": 102,
|
||||
"title": "Auth Health",
|
||||
"type": "row"
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 100 },
|
||||
{ "color": "red", "value": 500 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 0, "y": 35 },
|
||||
"id": 9,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Auth Attempts (1h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(auth_attempts_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[1h]))",
|
||||
"legendFormat": "attempts"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 5 },
|
||||
{ "color": "red", "value": 20 }
|
||||
]
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 8, "y": 35 },
|
||||
"id": 10,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Auth Failures (1h)",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(increase(auth_failures_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[1h]))",
|
||||
"legendFormat": "failures"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"color": { "mode": "palette-classic" },
|
||||
"thresholds": {
|
||||
"mode": "absolute",
|
||||
"steps": [
|
||||
{ "color": "green", "value": null },
|
||||
{ "color": "yellow", "value": 0.05 },
|
||||
{ "color": "red", "value": 0.2 }
|
||||
]
|
||||
},
|
||||
"unit": "percentunit"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 8, "x": 16, "y": 35 },
|
||||
"id": 11,
|
||||
"options": {
|
||||
"reduceOptions": { "calcs": ["lastNotNull"] },
|
||||
"colorMode": "value",
|
||||
"graphMode": "area"
|
||||
},
|
||||
"title": "Auth Failure Rate",
|
||||
"type": "stat",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(rate(auth_failures_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval])) / sum(rate(auth_attempts_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "failure rate"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"fieldConfig": {
|
||||
"defaults": {
|
||||
"custom": {
|
||||
"drawStyle": "line",
|
||||
"lineInterpolation": "smooth",
|
||||
"fillOpacity": 20,
|
||||
"stacking": { "mode": "none" }
|
||||
},
|
||||
"unit": "short"
|
||||
}
|
||||
},
|
||||
"gridPos": { "h": 8, "w": 24, "x": 0, "y": 43 },
|
||||
"id": 12,
|
||||
"title": "Auth Trends",
|
||||
"type": "timeseries",
|
||||
"targets": [
|
||||
{
|
||||
"expr": "sum(rate(auth_attempts_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "attempts/s"
|
||||
},
|
||||
{
|
||||
"expr": "sum(rate(auth_failures_total{service_name=~\"$service\", deployment_environment=~\"$env\"}[$__rate_interval]))",
|
||||
"legendFormat": "failures/s"
|
||||
}
|
||||
]
|
||||
}
|
||||
],
|
||||
"schemaVersion": 39,
|
||||
"tags": ["airi", "users", "auth", "grafana-cloud"],
|
||||
"templating": {
|
||||
"list": [
|
||||
{
|
||||
"current": { "selected": false, "text": "production", "value": "production" },
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"definition": "label_values(http_server_request_duration_milliseconds_count, deployment_environment)",
|
||||
"hide": 0,
|
||||
"includeAll": true,
|
||||
"multi": false,
|
||||
"name": "env",
|
||||
"query": "label_values(http_server_request_duration_milliseconds_count, deployment_environment)",
|
||||
"refresh": 2,
|
||||
"type": "query"
|
||||
},
|
||||
{
|
||||
"current": { "selected": false, "text": "server", "value": "server" },
|
||||
"datasource": { "type": "prometheus", "uid": "grafanacloud-projairi-prom" },
|
||||
"definition": "label_values(http_server_request_duration_milliseconds_count{deployment_environment=~\"$env\"}, service_name)",
|
||||
"hide": 0,
|
||||
"includeAll": true,
|
||||
"multi": true,
|
||||
"name": "service",
|
||||
"query": "label_values(http_server_request_duration_milliseconds_count{deployment_environment=~\"$env\"}, service_name)",
|
||||
"refresh": 2,
|
||||
"type": "query"
|
||||
}
|
||||
]
|
||||
},
|
||||
"time": { "from": "now-24h", "to": "now" },
|
||||
"title": "AIRI User & Auth (Cloud)",
|
||||
"uid": "airi-user-auth-cloud"
|
||||
}
|
||||
@@ -11,11 +11,15 @@ const tracer = trace.getTracer('airi-server-hono')
|
||||
|
||||
/**
|
||||
* Hono middleware that creates spans for each request and records
|
||||
* custom HTTP metrics (duration, active requests, status codes).
|
||||
* active request counts.
|
||||
*
|
||||
* NOTICE: Request duration is intentionally NOT recorded here. The
|
||||
* Node HTTP instrumentation already emits `http.server.request.duration`,
|
||||
* and recording the same metric here would double-count every request in
|
||||
* Grafana panels that read the histogram `_count` series as request rate.
|
||||
*/
|
||||
export function otelMiddleware(http: HttpMetrics): MiddlewareHandler<HonoEnv> {
|
||||
return async (c, next) => {
|
||||
const startTime = performance.now()
|
||||
const method = c.req.method
|
||||
const path = c.req.path
|
||||
|
||||
@@ -38,12 +42,6 @@ export function otelMiddleware(http: HttpMetrics): MiddlewareHandler<HonoEnv> {
|
||||
if (status >= 500) {
|
||||
span.setStatus({ code: SpanStatusCode.ERROR, message: `HTTP ${status}` })
|
||||
}
|
||||
|
||||
http.requestDuration.record((performance.now() - startTime) / 1000, {
|
||||
'http.request.method': method,
|
||||
'http.route': path,
|
||||
'http.response.status_code': status,
|
||||
})
|
||||
}
|
||||
catch (err) {
|
||||
const errorMessage = errorMessageFromUnknown(err)
|
||||
|
||||
Reference in New Issue
Block a user