Given that the timeout period on backend for an existing transaction
is 15 seconds, sleeping for 5 seconds between reconnect attempts seems
excessive. Lower the timer to 1 second, which should give us a slightly
better chance to avoid timeouts.
Change-Id: Ib74480f5630865cb7a11ca7027e0495443d1d14e
Signed-off-by: Robert Varga <robert.varga@pantheon.tech>
(cherry picked from commit
70f287502823bab284555b52b91043c3204b829b)
}
private static final Logger LOG = LoggerFactory.getLogger(ClientActorBehavior.class);
- private static final FiniteDuration RESOLVE_RETRY_DURATION = FiniteDuration.apply(5, TimeUnit.SECONDS);
+ private static final FiniteDuration RESOLVE_RETRY_DURATION = FiniteDuration.apply(1, TimeUnit.SECONDS);
/**
* Map of connections to the backend. This map is concurrent to allow lookups, but given complex operations