Hello All,
This is my first question into this mailing list and am looking for some
insights from more experienced Kafka users.
We are currently setting up a Apache Kafka cluster to create an event
based/stream processing system.
In our system we have a web based UI for doing certain operations manually,
similar to the standard automated processing.
We were thinking of tackling the UI like this.
1. UI HTTP request comes in, and gets written to a ‘request’ topic.
2. The standard processing gets the request from the ‘request’ topic and
processes it asynchronously.
3. Once the standard processing is done, the response is written to a
‘response’ topic.
4. We only return an HTTP response once we see the corresponding response on
the ‘response’ topic.
Basically:
normal automated producer ——> request topic ——> processor (kafka
consumer) ——> response topic
request HTTP ——> request topic
<—— wait for corresponding event on response topic
<—— respond HTTP
Mapping request response would be doing by a unique id (maybe offset).
Not sure if this is a good approach, right way of doing things.
All thoughts are welcome.
Regards,
Bruno Rassaerts | Freelance Java Developer
Novazone, Edingsesteenweg 302, B-1755 Gooik, Belgium
T: +32(0)54/26.02.03 - M: +32(0)477/39.01.15
[email protected] - www.novazone.be
