Is this so that a publisher can keep retrying publishing calls until it succeeds? This can be happen if the publisher/owner of event type observes the status codes for publish requests and fixes the event type (auth, schema etc) in time before retry policy expires so the failed calls can succeed. If not then this will just lead to meaning less calls to nakadi which nakadi will always reject, it occupies resources from client as well as from nakadi side.
I would like to understand the use case for which
nakad-java
client retries publishing even in case of 4xx errors.Is this so that a publisher can keep retrying publishing calls until it succeeds? This can be happen if the publisher/owner of
event type
observes the status codes for publish requests and fixes theevent type
(auth
,schema
etc) in time beforeretry policy
expires so the failed calls can succeed. If not then this will just lead to meaning less calls tonakadi
whichnakadi
will always reject, it occupies resources from client as well as fromnakadi
side.