Chiradeep Das
05/02/2026, 6:32 PMRocky
05/02/2026, 6:32 PMignore_json_size header, you need to modify the BE (Backend) configuration.
The specific parameter controlling this limit is streaming_load_max_batch_size_mb.
1. Update the BE Configuration File
To make the change permanent, modify the be.conf file on every Backend node:
1. Open be.conf (typically located in the conf/ directory of your BE installation).
2. Add or update the following line (for example, to set it to 500MB):
properties
streaming_load_max_batch_size_mb = 500
3. Restart the BE service for the changes to take effect.
──────────
2. Update Dynamically (No Restart Required)
Since this parameter is mutable, you can apply the change immediately without restarting the BE by using the update_config API. Run this command for each BE node in your cluster:
bash
curl -X POST http://<be_host>:<be_http_port>/api/update_config?streaming_load_max_batch_size_mb=500
_Replace <be_host> with the BE IP/hostname and <be_http_port> with the be_http_port (default is 8040)._
Summary of Relevant Parameters
* streaming_load_max_batch_size_mb: Specifically limits the size of JSON format data for Stream Loads. The default is 100 (MB).
* streaming_load_max_mb: Limits the overall file size for all Stream Loads (CSV, Parquet, etc.). The default is 102400 (100GB) in v3.0+. If your total JSON file is extremely large (e.g., >10GB), you may need to check this as well, though usually, it is already high enough.
Note: When increasing this limit, ensure your BE nodes have sufficient memory, as parsing very large JSON objects can be memory-intensive. Also, ensure the stream_load_default_timeout_second (an FE config) is long enough to accommodate the transfer of larger files.
References
• administration/management/BE_configuration.md
• faq/loading/Stream_load_faq.md
• sql-reference/sql-statements/loading_unloading/STREAM_LOAD.md
• loading/StreamLoad.md
• sql-reference/sql-statements/loading_unloading/BROKER_LOAD.md
• administration/management/FE_configuration.md