Correct optional fields can be omitted from the Heka message. 

The uuid and timestamp should remain required and set by the message creator, 
otherwise: 
- If the message creator was delayed in sending the messages to Heka when the 
backlog was processed all messages would basically be given the same timestamp. 
- If the creator sent duplicates of the same message it would be very difficult 
to correctly de-dupe without a uuid and timestamp. 

If you have your own protobuf format nothing prevents you from writing your own 
decoder and transforming the result into a Heka message. 

Trink 

----- Original Message -----

> From: "bruno binet" <[email protected]>
> To: "heka" <[email protected]>
> Sent: Tuesday, June 24, 2014 5:14:58 AM
> Subject: [heka] using the protobuf decoder

> Hi,

> I'd like to use heka to forward and process the data generated by an
> application. This is a new application so I'm free to choose any
> serialization format for data sent to Heka. Since Heka uses protobuf as the
> data format to communicate between multiple instances, I think this could be
> a good thing to go with protocol buffers.

> My question is: given the current message.proto file (
> https://github.com/mozilla-services/heka/blob/0ce27cb548df1fb31f3def56e3a9b6c1e28dcb77/message/message.proto
> ), many fields seem to be optional, so it means I don't have to set them
> (heka won't fail if these fields are not set)?

> And two fields are required in the "Message" message:
> required bytes uuid = 1 ;
> required int64 timestamp = 2 ; // nanoseconds since UNIX epoch
> I would like those fields to be optional so that heka could set a value for
> them if they are missing (it would set the timestamp to the current time and
> generate a new uuid).
> Woult it be possible?

> May be this does not make sense if the protocol buffer decoder is only meant
> to be used internally, but in this case, having a way to specify a custom
> protocol buffer decoder may be a good thing to have? (we could specify a
> custom proto file to load)

> Thanks,
> Bruno

> _______________________________________________
> Heka mailing list
> [email protected]
> https://mail.mozilla.org/listinfo/heka
_______________________________________________
Heka mailing list
[email protected]
https://mail.mozilla.org/listinfo/heka

Reply via email to