[
https://issues.apache.org/jira/browse/KAFKA-656?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14094589#comment-14094589
]
Jay Kreps commented on KAFKA-656:
---------------------------------
Hey [~ab10anand], I would rather not do explicit leases. The problem is that
zookeeper is a very scarce resource so you have to then quota the lease updates
themselves--basically it just seems complex. In any case, whether the mechanism
is leases or error, the client needs some mechanism to throttle itself when it
reaches the maximum of it's leased capacity or gets an error. I think the
errors can do this naturally. Our new java producer actually already has a
notion of a retry backoff, and I think we could generalize this to handle this
case fairly easily. This will cause the client to produce backpressure when it
hits its maximum.
> Add Quotas to Kafka
> -------------------
>
> Key: KAFKA-656
> URL: https://issues.apache.org/jira/browse/KAFKA-656
> Project: Kafka
> Issue Type: New Feature
> Components: core
> Affects Versions: 0.8.1
> Reporter: Jay Kreps
> Labels: project
>
> It would be nice to implement a quota system in Kafka to improve our support
> for highly multi-tenant usage. The goal of this system would be to prevent
> one naughty user from accidently overloading the whole cluster.
> There are several quantities we would want to track:
> 1. Requests pers second
> 2. Bytes written per second
> 3. Bytes read per second
> There are two reasonable groupings we would want to aggregate and enforce
> these thresholds at:
> 1. Topic level
> 2. Client level (e.g. by client id from the request)
> When a request hits one of these limits we will simply reject it with a
> QUOTA_EXCEEDED exception.
> To avoid suddenly breaking things without warning, we should ideally support
> two thresholds: a soft threshold at which we produce some kind of warning and
> a hard threshold at which we give the error. The soft threshold could just be
> defined as 80% (or whatever) of the hard threshold.
> There are nuances to getting this right. If you measure second-by-second a
> single burst may exceed the threshold, so we need a sustained measurement
> over a period of time.
> Likewise when do we stop giving this error? To make this work right we likely
> need to charge against the quota for request *attempts* not just successful
> requests. Otherwise a client that is overloading the server will just flap on
> and off--i.e. we would disable them for a period of time but when we
> re-enabled them they would likely still be abusing us.
> It would be good to a wiki design on how this would all work as a starting
> point for discussion.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
