[ci] Enable canton amplification improvements (#3151)

Signed-off-by: Nicu Reut <[email protected]>

Author: nicu-da

apps/app/src/test/resources/include/mediators.conf

@@ -15,6 +15,7 @@ _mediator_template {
   time-tracker.min-observation-duration = 10s
   sequencer-client {
     use-new-connection-pool = false
+    enable-amplification-improvements = true
   }
   topology.validate-initial-topology-snapshot = false
 }

apps/app/src/test/resources/include/participants.conf

@@ -45,6 +45,7 @@ _participant_template {
     # Use a higher number of in flight batches to increase throughput
     maximum-in-flight-event-batches = 50
     use-new-connection-pool = false
+    enable-amplification-improvements = true
   }
 
   # TODO(DACH-NY/canton-network-node#8331) Tune cache sizes

cluster/images/canton-mediator/app.conf

@@ -22,6 +22,7 @@ canton {
       sequencer-client {
         acknowledgement-interval = 10m
         use-new-connection-pool = false
+        enable-amplification-improvements = true
       }
 
       storage = ${_storage}

cluster/images/canton-participant/app.conf

@@ -30,6 +30,7 @@ canton {
         # Use a higher number of in flight batches to increase throughput
         maximum-in-flight-event-batches = 50
         use-new-connection-pool = false
+        enable-amplification-improvements = true
       }
 
       monitoring.grpc-health-server {

cluster/images/multi-participant/pre-bootstrap.sh

@@ -30,6 +30,7 @@ canton.participants.participant_$index = {
       # Use a higher number of in flight batches to increase throughput
       maximum-in-flight-event-batches = 50
       use-new-connection-pool = false
+      enable-amplification-improvements = true
     }
 
     monitoring.grpc-health-server {

docs/src/release_notes.rst

@@ -5,6 +5,15 @@
 
 .. _release_notes:
 
+Upcoming
+========
+
+- Sequencer connections
+
+    - Improve retries for sending sequencer submissions when a sequencer rejects the request with an overloaded error code by retrying immediately on another node.
+    - The network timeout for the connection was lowered to 15 seconds to detect failures faster.
+
+
 0.5.1
 =====