What the data for this category looks like
The data for this category primarily comes from advertising placement platforms, creative material libraries, user interaction data, and conversion statistics systems. Daily produced documents include creative scripts, material test sheets, placement data reports, and iteration checklists. Data fields cover creative material type, placement targeting, impressions, click-through rate, conversion cost, and more. Data update frequency is typically daily or weekly. Material combination data for single tests is synchronized in real time alongside placement progress. Test conclusions from batch iterations are summarized and archived on a weekly basis.
What constraints these characteristics impose on deployment and upgrades
The data sources for this category are scattered and have high update frequencies. Deployed services must use stable database and caching tools to ensure real-time synchronization of test data and placement logs. Daily documents include multiple types of structured tables and material files. Deployment must adapt to capacity and access efficiency requirements for file storage. Placement performance directly correlates with business metrics. Upgrade processes must strictly enforce service startup checks to avoid service interruptions caused by missing configurations. Multi-dimensional data from creative tests uses AI Proxy to call models and generate analysis conclusions. The two server environment variables AIPROXY_API_ENDPOINT (root address of the AI Proxy service, do not append /v1 or /chat/completions, not a single model address) and AIPROXY_API_TOKEN (AI Proxy administrator token, must match the AI Proxy server configuration and cannot be empty) must be configured correctly to prevent model call link failures.
How to configure settings
The following configurations follow official specifications for FastGPT v4.17.0.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
MONGODB_URI | mongodb://username:password@host:port/ad_creative_test?authSource=admin | Stores creative test data, placement logs, and iteration checklists. Requires specifying the authentication source to ensure data security |
REDIS_URL | redis://:your-redis-password@host:port/0 | Caches real-time synchronized user interaction data and placement metrics to reduce database query pressure |
DB_MAX_LINK | 100–200 connections | Adapts to concurrent creative test requests and avoids service freezes caused by exhausted connections |
STORAGE_S3_ENDPOINT | http://s3-cn-beijing.aliyuncs.com (adjust based on actual cloud service provider) | Stores creative material files and test reports. Adapts to large file upload and download requirements |
FE_DOMAIN | https://ad-creative-test.your-domain.com | Configures the frontend access domain for the creative test tool to ensure normal cross-domain requests between frontend and backend services |
ROOT_KEY | Randomly generated string of 32 or more characters with no special symbols | Serves as the highest-privilege authentication key for the system backend to ensure operational security during deployment |
Three common configuration mistakes
- Startup error messages indicating missing
AIPROXY_API_ENDPOINTorAIPROXY_API_TOKENtypically occur when these two required parameters are not configured in the system environment variables of the App/Admin server, or when they are mistakenly entered in model channel configurations, workflow global variables, application API request headers, or the standard model management page. - Frontend pages fail to load historical creative test data. This is caused by a mismatch between the domain configured in
FE_DOMAINand the actual access domain, resulting in blocked cross-domain requests. - Model calls return empty results or time out. Troubleshoot by verifying that
AIPROXY_API_ENDPOINTandAIPROXY_API_TOKENare configured correctly, and that model channel configurations are functioning normally.
How to confirm configurations are valid
- Check the system environment variables of the App/Admin server to confirm that
AIPROXY_API_ENDPOINT(root address of the AI Proxy service, do not append /v1 or /chat/completions, not a single model address) andAIPROXY_API_TOKEN(AI Proxy administrator token, must match the AI Proxy server configuration and cannot be empty) are correctly configured and not empty. - Run the service startup script and review startup logs for successful connection prompts for the database, cache, and file storage. No errors indicate basic configurations are normal.
- Upload a creative material file to confirm the file storage service can complete upload and download operations normally. This verifies the effectiveness of the
STORAGE_S3_ENDPOINTconfiguration. - Initiate a simulated creative test request to confirm the frontend page can normally load test data and model return results. This verifies that
DB_MAX_LINKandREDIS_URLconfigurations adapt to concurrent requirements.
The values provided are common starting points and should be evaluated against local samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-10-19.