What the data for this category looks like
Advertising and marketing data primarily comes from financial institution ad platform backends, marketing asset source storage repositories, and user interaction lead collection systems. Delivery data updates hourly or daily. Marketing asset files sync with marketing campaign cycles. User interaction data is generated in real time. Document structures include fields such as campaign ID, delivery channel, asset code, impressions, cost per click, and more. Units include impressions, yuan, seconds, and others. Asset files contain multi-format content such as financial product copy, poster source files, video clip segments, and more.
What constraints do these characteristics impose on workflow orchestration
The multi-source, heterogeneous data structure of advertising and marketing requires workflow configurations to include cross-data source field mapping nodes, to unify the format of data exported from different financial delivery platforms. Real-time updated delivery and interaction data requires workflows to support scheduled trigger or event trigger modes, to adapt to data synchronization rhythms. Multi-format marketing asset files require workflows to include parsing nodes adapted to different file types, to avoid parsing failures. Data isolation requirements across different marketing campaigns require workflow configuration of node-level permission filtering rules, to ensure compliance with financial customer data access regulations.
How to set the configurations
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
PARSE_FILE_TIMEOUT_SECONDS | 120–180 seconds | Financial marketing assets include product introduction videos and high-definition posters, with longer parsing durations than general documents |
maxContext | 8000–12000 characters | Financial product copy and marketing scripts are usually lengthy, requiring sufficient context for content analysis |
Recall Count | Top 8–12 entries | Marketing content related to a single marketing campaign is concentrated; excessive recall will introduce redundant associated data |
Workflow Trigger Mode | Dual mode: scheduled trigger + event trigger | Delivery data requires hourly synchronization, while user interaction data requires real-time response to financial lead generation needs |
TEXT2SQL_TABLE_WHITELIST | Delivery report table, asset library table, interaction lead collection table | Limit access only to data tables required for business, complying with financial data isolation requirements |
UPLOAD_FILE_MAX_SIZE | 500 MB | Financial marketing assets are mostly high-definition posters or short-duration product introduction videos, requiring adaptation to larger file uploads |
The parameter values provided on this page are general recommendations to serve as a starting point for configuration. Actual values are affected by material format, data volume, and business rules. Specific issues require individual analysis. It is recommended to test on your own samples before finalizing settings.
Three Common Mistakes
- Symptom: After importing a workflow from another environment, referenced custom plugins fail to load, and nodes display errors. Cause: The local storage path of the plugin was not bound during export. Only the workflow JSON configuration was copied, causing the new environment to fail to recognize the plugin's location.
- Symptom: Frequent timeout errors occur when parsing financial marketing assets. Cause: The
PARSE_FILE_TIMEOUT_SECONDSparameter was not adjusted. The default timeout duration for general documents was used, which cannot adapt to parsing long videos or high-definition assets. - Symptom: The associated content returned after workflow triggering exceeds the range that business processing can handle. Cause: The
Recall Countparameter was not set to a reasonable value. Excessive irrelevant historical marketing assets were recalled, resulting in redundant results.
How to Confirm the Configuration Is Complete
- Upload a single test financial marketing asset, verify that the structured fields output by the parsing node match expectations, and confirm that the parsing duration does not exceed the configured timeout threshold.
- Trigger a preset workflow trigger task, check whether the execution log includes correct cross-data source field mapping results, and confirm that cross-source data alignment takes effect.
- Export the workflow as a JSON file, then import it into a test environment. Check that all referenced plugin nodes have no missing errors, and confirm that the path configuration is correct.
- Call the TEXT2SQL node to generate a query statement, verify that only data tables within the whitelist can be accessed, and there are no prompts for unauthorized table access.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-14.