I completely agree that the typical use case of Redis is a distributed cache. That said, it also functions well as a temporary queue for event data en route to another data store. For example, Redis may be used to improve the fault-tolerance of a logging pipeline by serving as an intermediary en route to a separate storage layer (i.e. log4j2 -> Redis <- Logstash -> Elasticsearch).
Kafka is well-suited for these use cases, but Redis may be preferred if users want an in-memory broker rather than one that writes to disk. Kevin On 2018/08/07 15:24:31, Ralph Goers <[email protected]> wrote: > While I wouldn’t be opposed to a Redis Appender, I would find it a strange use of Redis. Normally, Redis is used as a distributed cache, not as a place to store log evens. Can you explain what the use case would be when using it?> > > Ralph> > > > On Aug 7, 2018, at 8:05 AM, Kevin Meurer <[email protected]> wrote:> > > > > > Hi all,> > > > > > I wanted to float the idea of folding a Redis log4j2 appender into the> > > log4j2 repository or releasing a canonical version under the Apache license.> > > > > > As you might be aware, one repository> > > <https://github.com/vy/log4j2-redis-appender> currently exists to meet this> > > need, but it is licensed under GPL and therefore not usable for a variety> > > of contexts. I think of a Redis appender as potentially analogous to the> > > existing Kafka Appender, as Redis is often used in a similar message broker> > > context.> > > > > > Is this something I could potentially add support for in log4j2? If not, is> > > there interest in an effort to create a separate repository to meet this> > > need?> > > > > > Any input would be appreciated.> > > > > > Thanks,> > > Kevin Meurer (kmeurer <https://github.com/kmeurer>)> > > >
