fix(infra): fix deploy tracking dashboard datasource UIDs and instant queries

- Add explicit uid: deploy-tracking to datasource provisioning
- Add instant: true to all Prometheus stat/gauge panel queries
- Pushgateway gauges need instant queries, not range queries

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

docker/grafana/dashboards/deploy-tracking.json

@@ -40,6 +40,7 @@
 			"targets": [
 				{
 					"expr": "time() - deploy_last_timestamp_seconds{branch=\"main\"}",
+					"instant": true,
 					"legendFormat": ""
 				}
 			],
@@ -78,6 +79,7 @@
 			"targets": [
 				{
 					"expr": "deploy_last_status{branch=\"main\"}",
+					"instant": true,
 					"legendFormat": ""
 				}
 			],
@@ -110,6 +112,7 @@
 			"targets": [
 				{
 					"expr": "deploy_last_duration_seconds{branch=\"main\"}",
+					"instant": true,
 					"legendFormat": ""
 				}
 			],
@@ -332,6 +335,7 @@
 			"targets": [
 				{
 					"expr": "deploy_service_image_size_mb",
+					"instant": true,
 					"legendFormat": "{{service}}"
 				}
 			],
@@ -379,6 +383,7 @@
 			"targets": [
 				{
 					"expr": "deploy_service_healthy",
+					"instant": true,
 					"legendFormat": "{{service}}"
 				}
 			],

docker/grafana/provisioning/datasources/deploy-tracking.yml

@@ -5,6 +5,7 @@ apiVersion: 1
 
 datasources:
   - name: DeployTracking
+    uid: deploy-tracking
     type: postgres
     access: proxy
     url: postgres:5432