What the Data for This Category Looks Like
This category of data includes hot metal and scrap charging data, L2 process control parameters, laboratory composition reports, MES production scheduling records, energy medium consumption data, and equipment operating status. Daily documents include heat plans, batching lists, converter smelting logs, continuous casting process parameter records, and slab quality judgment sheets. Data update frequencies cover per-heat, real-time collection, and shift-level summaries. Data fields include quantified metrics with units such as temperature, component proportion, and equipment load. The production process is high-temperature continuous, with strong coupling between heats and casting sequences, and production cannot be interrupted arbitrarily.
Constraints Imposed on Database and Operations Maintenance
The high real-time performance and multi-source heterogeneous data characteristics of this category impose multiple constraints on database operations and maintenance. High-frequency data generated per heat and in real time requires the database to support high-concurrency writes and low-latency queries. Multi-source data comes from different platforms such as L2 systems, laboratories, and MES, with varying formats and field standards, requiring flexible data structure adaptation. The production logic of strong coupling between heats and casting sequences requires data association queries to cover cross-table and temporal associations. Additionally, the production process cannot be shut down, requiring the database to support zero-downtime scaling and dynamic adjustment capabilities. The storage requirement for high-precision quantified metrics also requires the database to reserve sufficient field precision and storage space.
Configuration Guidelines
This page applies to FastGPT v4.17.0.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
MONGODB_URI | mongodb://fastgpt:securepass@db-host:27017/steel_trace?authSource=admin&readPreference=nearest | Adapts to high-concurrency writes for multi-source heterogeneous time-series data, and nearest-read reduces cross-region latency |
PG_URL | postgresql://fastgpt:securepass@pg-host:5432/fastgpt_mes?sslmode=require | Supports complex cross-heat and cross-casting sequence association queries to meet the association analysis needs of quality traceability |
DB_MAX_LINK | 200–300 connections | Matches the connection count requirements for per-heat real-time writes and concurrent queries across multiple departments, calibrated via actual testing |
REDIS_URL | redis://:securepass@redis-host:6379/0?maxmemory-policy=allkeys-lru&maxmemory=16 GB | Caches real-time heat status and high-frequency query results to reduce load pressure on core databases |
Three Common Misconfigurations
- Symptom: Database write requests pile up, disk read/write utilization remains high, and
connection timeouterrors appear in logs. Cause: The connection pool upper limit was not adjusted based on the high-frequency per-heat real-time write feature, leading to connection exhaustion and inability to process new requests. - Symptom: Associated data for some slab quality judgment sheets is missing, and query results return empty fields. Cause: Temporal association rules for multi-source data were not configured, leading to incorrect binding of heat data from different systems.
- Symptom: Disk read/write utilization remains high, storage space is rapidly depleted in a short period, and hard disk damage may occur in extreme scenarios. Cause: Historical data archiving policies were not enabled, and expired smelting and process records were not cleaned up in heat batches, leading to unrestricted accumulation of redundant data.
How to Verify Proper Configuration
- Run the database connection test script to verify connectivity for
MONGODB_URIandPG_URL, and confirm that the connection count can cover the preset concurrent request scale. - Import full associated data for a single heat to verify that cross-table association queries correctly return complete associated results for heats, casting sequences, and quality judgments.
- Check disk storage monitoring metrics to confirm that historical data archiving policies trigger according to preset rules, and storage space growth rate meets expectations.
- Simulate concurrent query requests across multiple departments to confirm that the
DB_MAX_LINKconfiguration supports concurrent access without connection timeout errors.
The values provided here are common starting points and should be measured against the reader's own samples.