What data for this category looks like
Data related to game revenue yield comes from game backend payment tracking logs, payment gateway callback records, and third-party game data analysis APIs.
Data update cadence has two parts: full synchronization of all previous day's statistical data each early morning, and hourly incremental updates for some real-time dashboards.
Data documents use structured table or JSON formats, and include fields such as game ID, server ID, statistical date, total daily revenue, daily paying users, average revenue per paying user, and daily revenue from core props. Units are uniformly yuan, integers, or date formats.
Each data entry corresponds to daily statistics for a single server. Total data volume grows linearly with the number of live game servers and the length of the statistical cycle.
Constraints on database and operations workflows
Multi-source data access requires format validation and duplicate data deduplication before processing, to prevent dirty data from entering the statistics pipeline.
Large batches of daily full sync data consume significant database write resources. Configure reasonable batch write thresholds to avoid connection timeouts.
High-frequency write demands from hourly incremental data require database connection pools to maintain sufficient concurrent connections, to prevent task interruptions from exhausted connections.
Frequent multi-dimensional queries for game data by server and date require partitioned indexes built on high-frequency query dimensions, to improve query efficiency.
Revenue data is sensitive. Configure permission controls and data desensitization rules to prevent data leaks.
Historical data must be retained for at least one year for period-over-period comparisons, which increases storage resource usage.
How to set configurations
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
MONGO_CONNECTION_POOL_SIZE | 20–30 | Matches the hourly incremental write demand of game data, balances concurrent connections and memory usage |
BATCH_IMPORT_SIZE | 800–1200 records | Adapts to batch imports of daily full game data, avoids single-write timeouts |
SYNC_TASK_TIMEOUT | 7200 seconds | Covers the time required for full game data sync, prevents task interruptions mid-run |
MILVUS_PARTITION_NUM | Set by server ID and statistical date | Partitions by high-frequency query dimensions of game data, improves vector retrieval speed |
DATA_ARCHIVE_DAYS | 365 days | Retains one year of historical data to meet period-over-period comparison needs for daily reports |
UPLOAD_DATA_MAX_SIZE | Calibrated to the actual size of full data | Adapts to the import limit for full game statistics data, prevents large file import failures |
The parameter values provided on this page are common starting points for configuration. Actual values are affected by data format, volume, and business rules. Specific issues require targeted analysis. It is recommended to test on your own samples before finalizing settings.
Three common configuration errors
- Phenomenon: Database connection returns
Failed to connect to 192.168.xx.xx.:1433or MongoDB connection error. Cause: Database connection pool parameters are not configured correctly, leading to concurrent connections exceeding the database's allowed limit. - Phenomenon: Shared pages become unresponsive after exceeding expected business concurrency, while local model server resources are not exhausted. Cause: Matching values for
MILVUS_PARTITION_NUMandBATCH_IMPORT_SIZEare not adjusted, leading to backlogs in the vector retrieval queue. - Phenomenon: Core prop revenue fields in imported game data are empty. Cause: Field mapping rules matching the game data structure are not configured, leading to failure to correctly extract target fields during parsing.
How to confirm proper configuration
- Run a small-batch game data import test, check that import logs have no connection timeout or parsing failure records, and verify that imported data fields match the source data.
- Simulate expected business concurrency to send query requests, confirm that all interface return status codes are within normal ranges, with no timeout errors.
- Check database index configurations, confirm that partitioned indexes or indexes have been created for high-frequency query dimensions of game data.
- View data archive task run logs, confirm that historical data has completed automatic archiving and retention per the required business cycle.
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.