ECO analytics dashboard not updating after BOM changes in Change Order workflow

PLM Oracle Agile PLM
, , , , , , , ,
1

We’re experiencing a persistent issue with our ECO analytics dashboard in Agile 9.3.4. After submitting Change Orders that update BOM structures, the dashboard metrics remain stale for 12-24 hours instead of refreshing within the expected 2-hour window.

The problem appears related to the Data Mart ETL job not triggering properly after Change Order workflow completion. We’ve verified the Change Orders are approved and released successfully, but the dashboard shows outdated part counts and change metrics. When we manually run the ETL job through the admin console, the dashboard updates immediately.

I’ve checked the workflow triggers and they appear configured correctly, but something is preventing automatic ETL execution. Has anyone encountered similar dashboard cache invalidation issues after BOM modifications? Our engineering team relies heavily on real-time change analytics for compliance reporting, and this delay is impacting decision-making.

2

Thanks for the suggestions. I checked the workflow post-conditions and the Data Mart notification handler is present. The agile.properties file shows realtime refresh is enabled. However, I noticed in the ETL logs that there are permission warnings related to the analytics service account accessing certain Change Order attributes. Could insufficient read permissions on specific ECO fields cause the ETL trigger to fail silently?

3

One more thing to check - dashboard cache invalidation settings. Even if the ETL runs successfully, the dashboard might be serving cached data. In Agile 9.3.4, there’s a known issue where the analytics cache doesn’t properly invalidate when Change Orders affect BOMs through indirect relationships. The cache timeout might be set too high. Look in the Analytics Admin console under Cache Management and verify the ECO-related cache entries have reasonable TTL values (we use 15 minutes max).

4

I worked on this exact scenario last quarter. Here’s a systematic approach to resolve your dashboard refresh issues:

1. Data Mart ETL Job Monitoring: First, verify the ETL trigger mechanism is functioning. Navigate to Admin Console > Analytics > ETL Configuration and check the ‘Change Event Triggers’ section. Ensure ‘ECO Workflow Completion’ is listed and enabled. Then examine the ETL execution logs at <AGILE_HOME>/agile/analytics/logs/etl_trigger.log for entries corresponding to your Change Order completion times. Look for entries like ‘Trigger received: ECO-12345 workflow completed’ followed by ‘ETL job initiated’. If triggers are logged but jobs aren’t starting, you have a scheduling service issue.

2. Change Order Workflow Triggers: The workflow must properly signal completion to the analytics layer. Edit your ECO workflow and verify the final approval node has a post-save event configured with these parameters:

  • Event Type: Workflow Status Change
  • Target Service: Analytics Data Mart
  • Action: Trigger Incremental Refresh
  • Scope: Include affected items and BOM changes

If this event handler is missing or misconfigured, the workflow completes successfully but analytics never receives the update notification.

3. Dashboard Cache Invalidation: Agile 9.3.4 has a specific bug (documented in Oracle Support Note 2847291.1) where BOM changes through Change Orders don’t properly invalidate dependent dashboard caches. The workaround involves two steps:

a) Reduce cache TTL for ECO-related metrics: In Analytics Admin > Cache Configuration, set these values:

  • eco.dashboard.cache.ttl=900 (15 minutes)
  • bom.change.cache.ttl=600 (10 minutes)
  • affected.items.cache.ttl=300 (5 minutes)

b) Enable forced cache refresh on ETL completion: Add this parameter to agile.properties:

agile.analytics.cache.invalidate.on.etl=true

4. Service Account Permissions: The analytics ETL service account needs comprehensive read access. Grant these privileges:

  • ECO base class: Read, Read All Attributes
  • Affected Items: Read, Read Redlines
  • BOM structure: Read, Read Effectivity
  • Manufacturer Parts: Read (if your dashboards include sourcing metrics)

5. Verification Steps: After implementing these changes:

  • Submit a test Change Order with BOM modifications
  • Monitor etl_trigger.log for immediate trigger receipt
  • Check ETL execution logs for job completion (should finish within 5-10 minutes for incremental refresh)
  • Verify dashboard updates within 15-20 minutes total
  • If still delayed, manually clear all analytics caches and retest

The root cause is typically a combination of missing workflow event handlers and overly aggressive cache settings. The 12-24 hour delay you’re seeing suggests you’re falling back to the nightly full ETL refresh rather than getting incremental updates. Implementing all three focus areas - proper ETL monitoring, correct workflow triggers, and appropriate cache invalidation - should restore your expected 2-hour refresh window.