Protobuf

The protobuf data format converts between JSON and binary protobuf wire format. It uses prost-reflect for dynamic message descriptors that the format compiles at runtime. The format requires no compile-time code generation. It ships as a separate crate, camel-dataformat-protobuf.

Marshal converts Body::Json to Body::Bytes. Unmarshal reverses the conversion and returns Body::Json. The round trip preserves field values through the JSON bridge.

When to use protobuf

Choose protobuf when the contract is a gRPC service or when the schema must evolve without breaking older clients. Protobuf carries typed fields, forward and backward compatibility, and a compact binary encoding. Choose JSON instead when the consumer is a browser, a REST API, or any system that reads text. JSON is readable, universal, and cheaper to debug. See Data Formats for the full format catalog.

Construction

ProtobufDataFormat takes a proto file path and a fully-qualified message name:

use camel_dataformat_protobuf::ProtobufDataFormat;

let df = ProtobufDataFormat::new("protos/helloworld.proto", "helloworld.HelloRequest")?;

The constructor compiles the proto file at runtime through camel-proto-compiler. Pass a shared ProtoCache to new_with_cache to reuse the compiled descriptor pool across formats.

The protobuf format is not built-in. Register it before the route starts:

ctx.data_format_registry()
    .register("protobuf", std::sync::Arc::new(df));

Body type support

DoS protection

The format rejects payloads larger than 64 MiB by default. The cap prevents out-of-memory errors from oversized inputs. Raise or lower the limit with with_max_decode_bytes:

let df = ProtobufDataFormat::new("schema.proto", "my.Message")?
    .with_max_decode_bytes(128 * 1024 * 1024);

Prost enforces a recursion limit of 100 levels. Deeply nested payloads return RecursionLimitReached at depth 100.

Reference: camel-dataformat-protobuf source