We migrated to Opcenter Cloud SOC 4.1 last month and performance analysis dashboards are now taking 45-60 seconds to load compared to 8-10 seconds on-premise. This affects real-time monitoring for production supervisors who need immediate visibility into line performance.
The issue started when we increased concurrent users from 50 to 120 after migration. With 50 users, dashboard performance was acceptable at 12-15 seconds. Now with 120 users, queries are timing out during peak production hours.
We’re using Azure SQL Database (Standard S3 tier) for the analytics database. Cloud infrastructure monitoring shows CPU utilization hitting 85-90% during dashboard loads and query execution times increasing proportionally with concurrent users.
Real-time monitoring is critical for identifying production bottlenecks and quality issues. Has anyone optimized performance analysis dashboards for high concurrent user loads in cloud deployments? We need to understand if this requires database autoscaling, query optimization, or implementing in-memory caching.
For performance dashboards, create composite indexes on PerformanceMetrics covering the most common query patterns. At minimum you need indexes on (ProductionDate, LineId) and (ShiftId, ProductionDate, LineId).
Also consider implementing Redis cache for dashboard data. Cache the aggregated metrics for 5-minute intervals so repeated dashboard loads hit cache instead of querying the database every time. This dramatically reduces database load for concurrent users viewing the same time periods.
Good points on both autoscaling and indexing. We haven’t added any custom indexes since migration - just using the default schema from Opcenter installation. What specific indexes would help most for typical performance dashboard queries that show hourly production rates and OEE calculations?
Autoscaling helps but won’t solve everything. The real issue is that Opcenter’s default dashboard queries aren’t optimized for cloud databases. They work fine on-premise with local SSD storage but struggle in cloud environments with network latency.
You need to add query optimization - specifically indexes on the PerformanceMetrics table for columns used in dashboard filters like ProductionDate, LineId, and ShiftId. Without proper indexes, every dashboard load does full table scans.