@kapa.ai in connector builder of airbyte 2.0.1, I have this error sometime it works fine in postman

Client error : 400 Bad Request {"exceptionStack":"OverflowError: int too big to convert\n\nThe above exception was the direct cause of the following exception:\n\nTraceback (most recent call last):\n  File \"/home/airbyte/.pyenv/versions/3.10.14/lib/python3.10/site-packages/airbyte_cdk/connector_builder/main.py\", line 82, in <module>\n    print(handle_request(sys.argv[1:]))\n  File \"/home/airbyte/.pyenv/versions/3.10.14/lib/python3.10/site-packages/airbyte_cdk/connector_builder/main.py\", line 77, in handle_request\n    return orjson.dumps(AirbyteMessageSerializer.dump(handle_connector_builder_request(source, command, config, catalog, state, limits))).decode()  # type: ignore[no-any-return] # Serializer.dump() always returns AirbyteMessage\nTypeError: Integer exceeds 64-bit range\n","exceptionClassName":"io.airbyte.protocol.models.AirbyteTraceMessage","message":"Error handling request: Integer exceeds 64-bit range"}

The knowledge sources don’t mention this exact

Integer exceeds 64-bit range

/

OverflowError: int too big to convert

error, but there are closely related issues in the Connector Builder: • Airbyte’s type system and JSON schemas assume integers fit in 64‑bit ranges. Values outside that range can cause failures in connectors and the Builder tooling.[Data types] • There is a known class of Builder bugs where certain values in records cause schema inference / serialization to crash (e.g. odd IDs being mis‑parsed, or large responses causing Builder server failures).[Large responses; Schema inferrer bug] In your case, Postman works because it just displays the raw HTTP response. The Connector Builder, however, must: 1. Parse the response, 2. Infer a JSON schema, 3. Serialize records as Airbyte messages. If any field is treated as an integer and its value is larger than a signed 64‑bit integer, the serialization step can raise exactly the error you’re seeing. Given what’s in the docs and forum, the practical workarounds you can try in the Builder are: 1. Force problematic fields to be strings in record processing In the Record processing / Transformations section of the Builder, cast any very large numeric IDs or counters to strings so they are not treated as integers by the schema/type system. This aligns with Airbyte’s supported types and avoids 64‑bit overflow.[Data types; Connector Builder overview] 2. Limit data during testing Similar to the schema‑inferrer issue where errors only appear with more records, try: ◦ Reducing page size / limit in your request, ◦ Testing with a smaller date or ID range, to identify which field/record introduces the huge integer.[Schema inferrer bug; Large responses] 3. Check for very large numeric fields Look at a failing response (e.g., via Postman) and search for: ◦ 64‑bit style IDs (very long numbers), ◦ Timestamps or counters represented as huge integers. Any such field is a candidate to be converted to string in the Builder. 4. If reproducible, open an issue Since similar Builder bugs are tracked on GitHub, if you can isolate a minimal response that triggers this error, it’s worth opening an issue with: ◦ Airbyte version (2.0.1), ◦ The failing endpoint and a redacted sample response, ◦ The full stack trace you posted.[Schema inferrer bug; Complex ID bug] Because the sources don’t document this exact stack trace, I can’t point to a confirmed fix, but based on the documented type limits and similar Builder issues, treating the offending large numeric fields as strings in your connector definition is the safest immediate workaround. 👍 0 button 👎 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.