Kafka Collective,
Since KSQL came out there is something I have been watching for in the
form of a blog post, article, or somewhere, but I have yet to see it. I may
have missed it over the last year+, so I figured I would check in and see
if anyone has resources. Here are my questions.
1. How are people pushing KSQL into production? I am guessing not by
opening a CLI and copy and paste? CRD's and some operator pattern?
2. How/can you deal with canary tests when pushing new versions?
3. How do you keep track of the current versions of the streams running?
4. How is telemetry handled? Just JMX cookie-cutter metrics? I assume there
is enough there to get stream by stream and table by table metrics? If they
are running, etc?
5. How is production-level error handling been for people? Dead letter
topics? skipping poison messages, etc?
Anything I missed as far as traditional best practices pushing things into
production? I see lots of cool demos, lots of examples, I use it for dev
work as I design what I want, but haven't heard or seen production level
KSQL and any good articles around the good, bad, and ugly, lessons learned,
etc.
Thanks,
Carl
