Hi Mehran,

Thanks for the input. We have talked about this quite a bit. You can find
references to the conversations in both the Reliable Logs Feature Narrative
<https://docs.google.com/document/d/1AhNDiuZef0rlNB9aCbBO-0DbBz_bC6iixHaWdjtBIHY/edit?usp=sharing>,
and the attached Logging Reliability white paper.

This is actually a challenging problem to solve. It's not clear where we
would enforce the quotas. Our current plans are to implement Line #2
in the Loggregator
Roadmap
<https://docs.google.com/spreadsheets/d/1QOCUIlTkhGzVwfRji7Q14vczqkBbFGkiDWrJSKdRLRg/edit#gid=0>
to
change the protocol between Diego and Metron to TCP. This will allow us to
track explicitly the discarded messages from a high logging volume app, and
notify the operator about the app. Once we gain some experience on the
actual behavior of apps in the wild, and with customer feedback, we'll
decide on the next steps.

Thanks,
Jim, Loggregator PM

On Thu, Jul 14, 2016 at 1:46 PM, Mehran Saliminia <msaliminia(a)gmail.com>
wrote:

Hi all,

I had a question regarding the number of logs that each app or app
instance in CF can generate. AFAIK there is no configuration option to
define a limit for maximum number of logs that each app is allowed to
generate per second.

Why don't you limit the log generation rate per app? Is there any
architectural/technical reason for that? or you don't see any need to have
such a rate limitter.

Regards,
Mehran

--
Jim Campbell | Product Manager | Cloud Foundry | Pivotal.io | 303.618.0963