diff --git a/cmd/frostfs-node/cache.go b/cmd/frostfs-node/cache.go
index 57f65d873..06142a46c 100644
--- a/cmd/frostfs-node/cache.go
+++ b/cmd/frostfs-node/cache.go
@@ -165,13 +165,11 @@ type ttlContainerStorage struct {
 	delInfoCache   *ttlNetCache[cid.ID, *container.DelInfo]
 }
 
-func newCachedContainerStorage(v container.Source, ttl time.Duration) ttlContainerStorage {
-	const containerCacheSize = 100
-
-	lruCnrCache := newNetworkTTLCache(containerCacheSize, ttl, func(id cid.ID) (*container.Container, error) {
+func newCachedContainerStorage(v container.Source, ttl time.Duration, containerCacheSize uint32) ttlContainerStorage {
+	lruCnrCache := newNetworkTTLCache(int(containerCacheSize), ttl, func(id cid.ID) (*container.Container, error) {
 		return v.Get(id)
 	}, metrics.NewCacheMetrics("container"))
-	lruDelInfoCache := newNetworkTTLCache(containerCacheSize, ttl, func(id cid.ID) (*container.DelInfo, error) {
+	lruDelInfoCache := newNetworkTTLCache(int(containerCacheSize), ttl, func(id cid.ID) (*container.DelInfo, error) {
 		return v.DeletionInfo(id)
 	}, metrics.NewCacheMetrics("container_deletion_info"))
 
diff --git a/cmd/frostfs-node/config.go b/cmd/frostfs-node/config.go
index 63f410b89..17a1cdc76 100644
--- a/cmd/frostfs-node/config.go
+++ b/cmd/frostfs-node/config.go
@@ -570,6 +570,8 @@ type cfgMorph struct {
 	// TTL of Sidechain cached values. Non-positive value disables caching.
 	cacheTTL time.Duration
 
+	containerCacheSize uint32
+
 	proxyScriptHash neogoutil.Uint160
 }
 
diff --git a/cmd/frostfs-node/config/morph/config.go b/cmd/frostfs-node/config/morph/config.go
index 1c536a0e2..d089870ea 100644
--- a/cmd/frostfs-node/config/morph/config.go
+++ b/cmd/frostfs-node/config/morph/config.go
@@ -30,6 +30,9 @@ const (
 
 	// FrostfsIDCacheSizeDefault is a default value of APE chain cache.
 	FrostfsIDCacheSizeDefault = 10_000
+
+	// ContainerCacheSizeDefault represents the default size for the container cache.
+	ContainerCacheSizeDefault = 100
 )
 
 var errNoMorphEndpoints = errors.New("no morph chain RPC endpoints, see `morph.rpc_endpoint` section")
@@ -103,6 +106,18 @@ func CacheTTL(c *config.Config) time.Duration {
 	return CacheTTLDefault
 }
 
+// ContainerCacheSize returns the value of "container_cache_size" config parameter
+// from "morph" section.
+//
+// Returns 0 if the value is not positive integer.
+// Returns ContainerCacheSizeDefault if the value is missing.
+func ContainerCacheSize(c *config.Config) uint32 {
+	if c.Sub(subsection).Value("container_cache_size") == nil {
+		return ContainerCacheSizeDefault
+	}
+	return config.Uint32Safe(c.Sub(subsection), "container_cache_size")
+}
+
 // SwitchInterval returns the value of "switch_interval" config parameter
 // from "morph" section.
 //
diff --git a/cmd/frostfs-node/container.go b/cmd/frostfs-node/container.go
index 5a29aac76..dc326c66a 100644
--- a/cmd/frostfs-node/container.go
+++ b/cmd/frostfs-node/container.go
@@ -86,44 +86,47 @@ func configureEACLAndContainerSources(c *cfg, client *cntClient.Client, cnrSrc c
 		cnrRdr.src = cnrSrc
 		cnrRdr.lister = client
 	} else {
-		// use RPC node as source of Container contract items (with caching)
-		cachedContainerStorage := newCachedContainerStorage(cnrSrc, c.cfgMorph.cacheTTL)
-		cachedEACLStorage := newCachedEACLStorage(eACLFetcher, c.cfgMorph.cacheTTL)
+		cachedContainerStorage := cnrSrc
+		if c.cfgMorph.containerCacheSize > 0 {
+			// use RPC node as source of Container contract items (with caching)
+			cachedContainerStorage := newCachedContainerStorage(cnrSrc, c.cfgMorph.cacheTTL, c.cfgMorph.containerCacheSize)
 
-		subscribeToContainerCreation(c, func(e event.Event) {
-			ev := e.(containerEvent.PutSuccess)
+			subscribeToContainerCreation(c, func(e event.Event) {
+				ev := e.(containerEvent.PutSuccess)
 
-			// read owner of the created container in order to update the reading cache.
-			// TODO: use owner directly from the event after neofs-contract#256 will become resolved
-			//  but don't forget about the profit of reading the new container and caching it:
-			//  creation success are most commonly tracked by polling GET op.
-			cnr, err := cnrSrc.Get(ev.ID)
-			if err == nil {
-				cachedContainerStorage.containerCache.set(ev.ID, cnr, nil)
-			} else {
-				// unlike removal, we expect successful receive of the container
-				// after successful creation, so logging can be useful
-				c.log.Error(logs.FrostFSNodeReadNewlyCreatedContainerAfterTheNotification,
+				// read owner of the created container in order to update the reading cache.
+				// TODO: use owner directly from the event after neofs-contract#256 will become resolved
+				//  but don't forget about the profit of reading the new container and caching it:
+				//  creation success are most commonly tracked by polling GET op.
+				cnr, err := cnrSrc.Get(ev.ID)
+				if err == nil {
+					cachedContainerStorage.containerCache.set(ev.ID, cnr, nil)
+				} else {
+					// unlike removal, we expect successful receive of the container
+					// after successful creation, so logging can be useful
+					c.log.Error(logs.FrostFSNodeReadNewlyCreatedContainerAfterTheNotification,
+						zap.Stringer("id", ev.ID),
+						zap.Error(err),
+					)
+				}
+
+				c.log.Debug(logs.FrostFSNodeContainerCreationEventsReceipt,
 					zap.Stringer("id", ev.ID),
-					zap.Error(err),
 				)
-			}
+			})
 
-			c.log.Debug(logs.FrostFSNodeContainerCreationEventsReceipt,
-				zap.Stringer("id", ev.ID),
-			)
-		})
+			subscribeToContainerRemoval(c, func(e event.Event) {
+				ev := e.(containerEvent.DeleteSuccess)
+				cachedContainerStorage.handleRemoval(ev.ID)
+				c.log.Debug(logs.FrostFSNodeContainerRemovalEventsReceipt,
+					zap.Stringer("id", ev.ID),
+				)
+			})
+		}
 
-		subscribeToContainerRemoval(c, func(e event.Event) {
-			ev := e.(containerEvent.DeleteSuccess)
-			cachedContainerStorage.handleRemoval(ev.ID)
-			c.log.Debug(logs.FrostFSNodeContainerRemovalEventsReceipt,
-				zap.Stringer("id", ev.ID),
-			)
-		})
-
-		c.cfgObject.eaclSource = cachedEACLStorage
 		c.cfgObject.cnrSource = cachedContainerStorage
+		cachedEACLStorage := newCachedEACLStorage(eACLFetcher, c.cfgMorph.cacheTTL)
+		c.cfgObject.eaclSource = cachedEACLStorage
 
 		cnrRdr.lister = client
 		cnrRdr.eacl = c.cfgObject.eaclSource
diff --git a/cmd/frostfs-node/morph.go b/cmd/frostfs-node/morph.go
index 7178cd97d..1bfcb8ac9 100644
--- a/cmd/frostfs-node/morph.go
+++ b/cmd/frostfs-node/morph.go
@@ -90,6 +90,7 @@ func initMorphComponents(ctx context.Context, c *cfg) {
 
 	var netmapSource netmap.Source
 
+	c.cfgMorph.containerCacheSize = morphconfig.ContainerCacheSize(c.appCfg)
 	c.cfgMorph.cacheTTL = morphconfig.CacheTTL(c.appCfg)
 
 	if c.cfgMorph.cacheTTL == 0 {
diff --git a/config/example/node.yaml b/config/example/node.yaml
index 86be35ba8..2a80fba18 100644
--- a/config/example/node.yaml
+++ b/config/example/node.yaml
@@ -81,6 +81,7 @@ morph:
   cache_ttl: 15s  # Sidechain cache TTL value (min interval between similar calls). Negative value disables caching.
                   # Default value: block time. It is recommended to have this value less or equal to block time.
                   # Cached entities: containers, container lists, eACL tables.
+  container_cache_size: 100 # container_cache_size is is the maximum number of containers in the cache.
   switch_interval: 3m # interval b/w RPC switch attempts if the node is connected not to the highest priority node
   rpc_endpoint:  # side chain NEO RPC endpoints; are shuffled and used one by one until the first success
     - address: wss://rpc1.morph.frostfs.info:40341/ws