As Jason wrote, this won't scale as the number of partitions increases. We already have users who have tens of thousands of topics, or more. If you multiply that by 100x over the next few years, you end up with this API returning full information about millions of topics, which clearly doesn't work.
We discussed this a lot in the original KIP-117 DISCUSS thread which added the Java AdminClient. ListTopics and DescribeTopics were deliberately kept separate because we understood that eventually a single RPC would not be able to return information about all the topics in the cluster. So I have to vote -1 for this proposal as it stands. I do agree that adding a way to describe topics by a regular expression on the server side would be very useful. This would also fix a major scalability problem we have now, which is that when subscribing via a regular expression, clients need to fetch the full list of all topics in the cluster and filter locally. I think a regular expression library like re2 would be ideal for this purpose. re2 is standardized and language-agnostic (it's not tied only to Java). In contrast, Java regular expression change with different releases of the JDK (there were some changes in java 8, for example). Also, re2 regular expressions are linear time, never exponential time. See https://github.com/google/re2j regards, Colin On Fri, Jul 13, 2018, at 05:00, Andras Beni wrote: > The KIP looks good to me. > However, if there is willingness in the community to work on metadata > request with patterns, the feature proposed here and filtering by '*' or > '.*' would be redundant. > > Andras > > > > On Fri, Jul 13, 2018 at 12:38 AM Jason Gustafson <ja...@confluent.io> wrote: > > > Hey Manikumar, > > > > As Kafka begins to scale to larger and larger numbers of topics/partitions, > > I'm a little concerned about the scalability of APIs such as this. The API > > looks benign, but imagine you have have a few million partitions. We > > already expose similar APIs in the producer and consumer, so probably not > > much additional harm to expose it in the AdminClient, but it would be nice > > to put a little thought into some longer term options. We should be giving > > users an efficient way to select a smaller set of the topics they are > > interested in. We have always discussed adding some filtering support to > > the Metadata API. Perhaps now is a good time to reconsider this? We now > > have a convention for wildcard ACLs, so perhaps we can do something > > similar. Full regex support might be ideal given the consumer's > > subscription API, but that is more challenging. What do you think? > > > > Thanks, > > Jason > > > > On Thu, Jul 12, 2018 at 2:35 PM, Harsha <ka...@harsha.io> wrote: > > > > > Very useful. LGTM. > > > > > > Thanks, > > > Harsha > > > > > > On Thu, Jul 12, 2018, at 9:56 AM, Manikumar wrote: > > > > Hi all, > > > > > > > > I have created a KIP to add describe all topics API to AdminClient . > > > > > > > > https://cwiki.apache.org/confluence/display/KAFKA/KIP- > > > 327%3A+Add+describe+all+topics+API+to+AdminClient > > > > > > > > Please take a look. > > > > > > > > Thanks, > > > > >
