diff --git a/journal/src/main/scala/com/evolutiongaming/kafka/journal/ConsumerPool.scala b/journal/src/main/scala/com/evolutiongaming/kafka/journal/ConsumerPool.scala
index 0ddc5214d..711a7fa26 100644
--- a/journal/src/main/scala/com/evolutiongaming/kafka/journal/ConsumerPool.scala
+++ b/journal/src/main/scala/com/evolutiongaming/kafka/journal/ConsumerPool.scala
@@ -26,10 +26,8 @@ private[journal] object ConsumerPool {
     for {
       cores <- Runtime[F].availableCores.toResource
       pool <- consumer.toResourcePool(
-        (cores.toDouble * poolConfig.multiplier)
-          .round
-          .toInt,
-        poolConfig.idleTimeout,
+        maxSize               = math.max(1, (cores.toDouble * poolConfig.multiplier).intValue),
+        expireAfter           = poolConfig.idleTimeout,
         discardTasksOnRelease = true,
       )
     } yield {
diff --git a/journal/src/main/scala/com/evolutiongaming/kafka/journal/Journal.scala b/journal/src/main/scala/com/evolutiongaming/kafka/journal/Journal.scala
index fa150950a..f714d0716 100644
--- a/journal/src/main/scala/com/evolutiongaming/kafka/journal/Journal.scala
+++ b/journal/src/main/scala/com/evolutiongaming/kafka/journal/Journal.scala
@@ -325,6 +325,15 @@ object Journal {
     implicit val configReaderCallTimeThresholds: ConfigReader[CallTimeThresholds] = deriveReader[CallTimeThresholds]
   }
 
+  /**
+   * Configuration for the dynamic pool of Kafka consumers used on recovery in case the data is not yet replicated to
+   * Cassandra
+   *
+   * @param multiplier defines max pool size = multiplier x number of cores,
+   *                   if a calculated max pool size is below 1, the size of 1 is used to avoid starting with
+   *                   an unusable consumer pool configuration
+   * @param idleTimeout if idle for this time, Kafka consumers are closed
+   */
   final case class ConsumerPoolConfig(
     multiplier: Double,
     idleTimeout: FiniteDuration,
@@ -339,12 +348,13 @@ object Journal {
 
   final case class DataIntegrityConfig(
     /**
-      * If true then duplicated [[SeqNr]] in events will cause [[JournalError]] `Data integrity violated`
+      * On recovery, if true, duplicated [[SeqNr]] in events will cause [[JournalError]] `Data integrity violated`
       */
     seqNrUniqueness: Boolean,
 
     /**
-      * If true then events with [[RecordId]] different from one in metadata will be filtered out
+      * On recovery, if true, events with [[RecordId]] different from the one in the current metadata record
+      * will be filtered out and logged as an error.
       */
     correlateEventsWithMeta: Boolean,
   )
diff --git a/persistence/src/main/resources/reference.conf b/persistence/src/main/resources/reference.conf
index 2d9a16805..f5babccd0 100644
--- a/persistence/src/main/resources/reference.conf
+++ b/persistence/src/main/resources/reference.conf
@@ -53,6 +53,28 @@ evolutiongaming.kafka-journal.persistence {
     }
 
     json-codec = default
+
+    # Configuration for the dynamic pool of Kafka consumers used on recovery in case the data is not yet replicated to
+    # Cassandra
+    consumer-pool {
+      # Defines max pool size = multiplier x number of cores.
+      # If a calculated max pool size is below 1, the size of 1 is used to avoid starting with an unusable consumer pool
+      # configuration.
+      # Multiplier value could be fractional.
+      multiplier = 10
+
+      # if idle for this time, Kafka consumers are closed
+      idle-timeout = 1min
+    }
+
+    data-integrity {
+      # On recovery, if true, duplicated sequence numbers in events will cause a JournalError `Data integrity violated`
+      seq-nr-uniqueness = true
+
+      # On recovery, if true, events with record IDs different from the one in the current metadata record
+      # will be filtered out and logged as an error
+      correlate-events-with-meta = false
+    }
   }
 
   dispatcher {