[zeromq-dev] Exploring 0mq 3.0 - lack of monitoring

[Pieter Hintjens]

On Sat, Oct 29, 2011 at 6:45 PM, Ilja Golshtein <ilejncs at narod.ru> wrote:

> Not sure I understand how it is possible to suggest HWM value regardless of application specific.

In the simplest case HWM works exactly as TCP buffers. Backpressure.
That's a good default situation that handles 80% of cases.

> Could you suggest better interface than getsockopt?
> Any ideas how to handle multiple peers (at least from API standpoint)?

To both of these questions: no. The API does not expose peer-specific
information. That is a good thing IMO.

What I'd do is perhaps:

* use CBFC where the simple defaults aren't the right design
* add internal monitoring (sys://log) to report on HWM overflows,
which can then be classified as CBFC failures
* use explicit app-level queues where it makes sense, and monitor those myself

-Pieter