What data for this category looks like
Data sources for this category include hot metal and scrap metal weighing data, L2 process control real-time data, laboratory element test reports, MES production execution data, energy medium consumption records, and equipment status monitoring data. Daily documents include heat plan sheets, batching ratio sheets, converter smelting process records, molten steel composition test reports, continuous casting process parameter records, and slab quality judgment sheets. Data update rhythm follows individual heats, real-time process nodes, and shift scheduling cycles. Fields include heat number, hot metal weight (tons), smelting duration (minutes), continuous casting pulling speed (meters per minute), slab thickness (millimeters), and others. There is no unified fixed field template.
What constraints these characteristics impose on the workflow orchestration link
This category has multi-source heterogeneous data with significantly different update rhythms. Workflow orchestration must split data pulling links by trigger nodes to avoid excessive load caused by pulling data from multiple systems synchronously. Heat and casting sequences are strongly coupled. Workflows must support binding processing flows by heat-casting sequence association, and cannot split individual processing links separately. High-temperature continuous production processes cannot be stopped and retried at will. Workflows must be configured with automatic failure rollback and breakpoint continuation mechanisms to ensure process continuity. Multiple types of documents have no unified format. Workflows must integrate text parsing links adapted to different field structures, while limiting the amount of data processed per time to avoid timeouts.
How to configure the settings
The following configurations are based on FastGPT v4.17.0.
| Configuration Item | Recommended Approach | Rationale |
|---|---|---|
variables | Defined by heat number + smelting timestamp, with a value range of 16-character strings | Uniquely identifies full-link data for a single heat, avoiding data confusion across heats |
nodeId | Configured by link type + sequence number, in the format link_name_two-digit sequence number | Quickly locate each processing node in the workflow to facilitate troubleshooting and process adjustments |
| Text parsing adaptation rules | Preset field mappings by document type, supporting 6 types of business documents including heat plans and batching sheets | Adapts to field extraction requirements for multiple types of heterogeneous documents in this category |
| Failure retry count | 2-3 times | Balances process stability and system resource consumption, avoiding infinite retries that occupy resources |
| Timeout threshold | 600 seconds | Matches the average duration of single-heat data processing, avoiding timeouts caused by data volume fluctuations |
| Single-process data volume | 1000–1500 MB, calibrated by actual testing | Matches the average size of single-heat data in this category, avoiding exceeding system processing limits |
Three common errors
- Phenomenon: Fields are empty when calling
variablesin the workflow, and the front-end interface displays a variable unbound prompt. Cause: The value range of global variables was not correctly bound before workflow startup, leading to lost data transfer across links. - Phenomenon: The HTTP request node returns status code 500, and logs show incompatible file formats. Cause: The correct file transfer format was not configured, and TXT files were not transferred to the backend for processing as binary streams.
- Phenomenon: After importing a workflow shared by others, the text processing module cannot be found in the local workflow. Cause: The enable switch for the corresponding function plugin was not turned on in the local workspace, causing the module to fail to load.
How to confirm the configuration is correct
- Start a test workflow, pass simulated heat data, and check whether
variablesare correctly bound to the heat number and timestamp. - Trigger the HTTP request node, verify that TXT format document data can be correctly transferred and the processed TXT file is returned.
- Check whether the
nodeIdof each workflow node matches the configuration list to ensure accurate node positioning. - Simulate data volume fluctuation scenarios, verify whether the timeout threshold and retry count configuration can cover abnormal situations.
The values provided are common starting points and should be evaluated against the reader's own samples.