What Data Looks Like for This Category
This category of data includes packaging process cards, bonding and molding parameters, wafer incoming material maps, test program versions, yield reports, failure analysis reports and other documents. Data sources cover wafer test maps, MES systems, bonding and molding equipment, ATE test machines, X-ray/SAM detection equipment and customer specification requirements. Data is updated per batch of materials, per individual chip test, and in daily summaries. Fields include wafer number, batch number, test bin code, equipment operating parameters, failure location and mode, among others. Some documents include binary image data.
Constraints Imposed on the Database and Operations Link
Multi-source heterogeneous data sources create format compatibility requirements. The system must adapt to structured MES data, binary wafer test maps and device logs. The high-frequency update rhythm requires the database to support high-concurrency writes and low-latency queries, to prevent data backlog from affecting traceability efficiency. The strong association between batches and individual chips requires table structure designs to include unique identifier fields, to ensure accurate data alignment. The growing share of binary image data requires configuring sufficient storage capacity and a reasonable indexing strategy. Data errors may allow defective chips to enter finished products, so supporting data verification and backup mechanisms is necessary to ensure data accuracy and recoverability.
Configuration Settings
These configurations are adapted for FastGPT v4.17.0.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
PG_URL | postgresql://{username}:{password}@{host}:5432/seal_pack_db?sslmode=require | Stores structured batch, yield and failure analysis data, supports transactional operations to ensure data consistency |
MONGODB_URI | mongodb://{username}:{password}@{host}:27017/seal_pack_img?retryWrites=true | Stores unstructured data such as binary wafer test maps and X-ray/SAM detection images, supports flexible field expansion |
DB_MAX_LINK | 200–300 connections | Adapts to high-frequency data writes and queries per batch and per chip, avoids connection exhaustion impacting business operations |
REDIS_URL | redis://:{password}@{host}:6379/0 | Caches frequently accessed test program versions and device parameters, reduces repeated query load on the database |
Three Common Configuration Errors
- The symptom is a
tool_callserror with a400status code returned during tool calling, or no database connection output in the workflow. The cause is incorrect configuration ofPG_URLorMONGODB_URIconnection parameters, or a mismatch between input variables and database field format requirements. - The symptom is successful manual SQL execution, but an error occurs when passing variables. The cause is failure to escape variables, leading to SQL syntax errors, or a mismatch between variable type and database field definition.
- The symptom is inability to call database tools in the deployment environment. The cause is that the container network configuration does not open database ports, or the authentication parameters for
REDIS_URLare configured incorrectly, leading to failed cache service connection.
How to Verify Correct Configurations
- Run the FastGPT built-in database connection test tool, verify the connectivity of
PG_URLandMONGODB_URI, confirm there are no connection timeout or authentication failure prompts. - Import a simulated packaging batch data set, call the database query interface via a workflow, verify that data can be read and written normally.
- Check the database connection pool monitoring, confirm that the current number of connections does not exceed the
DB_MAX_LINKconfiguration value, with no connection exhaustion. - Test the variable input SQL execution flow, verify that different types of variables (such as batch numbers, bin codes) can be correctly mapped to database query conditions.
The values provided are common starting points and should be measured against the reader's own samples.