Hi all, Thanks for the comments so far. Does anyone else have thoughts on either proposal? For middleware, I can put up an implementation draft if that's more useful. I could also include an example OpenTracing integration, though I'm not sure how to structure "contrib" modules (presumably we don't want that as an actual dependency).
Thanks, David On 7/8/19, Wes McKinney <wesmck...@gmail.com> wrote: > hi David, > > On the DoPut change, I am potentially in favor of making a change to > support the use case you are describing. In principle, the CMD field > of FlightDescriptor could be used to transmit application-defined > information with a Put request. Though I suspect that a DoAction to > "prepare" the server for a DoPut would be more common in production > servers. I'm interested in Jacques's or others' thoughts about this > > On the middleware API I am in favor; there are obviously a lot details > that have to be worked out, but it's important that Flight servers and > clients are able to be instrumented to collect metrics in a production > setting. > > Thanks, > Wes > > On Mon, Jul 8, 2019 at 9:33 AM David Li <li.david...@gmail.com> wrote: >> >> Hi all, >> >> I've put together two more proposals for Flight, motivated by projects >> we've been working on. I'd appreciate any comments on the >> design/reasoning; I'm already working on the implementation, alongside >> some other improvements to Flight. >> >> The first is to modify the DoPut call to follow the same request >> pattern as DoGet. This is a format change and would require a vote. >> >> https://docs.google.com/document/d/1hrwxNwPU1aOD_1ciRUOaGeUCyXYOmu6IxxCfY6Stj6w/edit?usp=sharing >> >> --- >> >> The second is to introduce a middleware API for Flight, to make >> integration with logging and tracing frameworks (e.g. OpenTracing) >> easier: >> >> https://docs.google.com/document/d/1Nm1ASp6w5PZQX42KkOTLK5LvrOv5q2cCbkaiVjqKjLU/edit?usp=sharing >> >> As part of this, I'm working on APIs to send richer error >> codes/messages between Flight clients/servers, now that @Micah has >> enabled custom error codes in Arrow-C++. These are API changes and >> should not require a vote, I believe. >> >> --- >> >> Anyone should have comment access to these documents, but let me know if >> not. >> >> Thanks! >> >> Best, >> David >