feat: enable vtprotobuf, watch batching, more limits

Fixes #5

Batch watch responses in a single batch so that client can quickly know
that initial snapshot got delivered.

Bump go.mod deps.

Implement more limits.

Signed-off-by: Andrey Smirnov <andrey.smirnov@talos-systems.com>

.codecov.yml

@@ -1,6 +1,6 @@
 # THIS FILE WAS AUTOMATICALLY GENERATED, PLEASE DO NOT EDIT.
 #
-# Generated on 2021-09-22T21:32:59Z by kres 2a27963-dirty.
+# Generated on 2021-09-23T14:58:27Z by kres 2a27963-dirty.
 
 codecov:
   require_ci_to_pass: false
@@ -9,7 +9,7 @@ coverage:
   status:
     project:
       default:
-        target: 15%
+        target: 30%
         threshold: 0.5%
         base: auto
         if_ci_failed: success

.kres.yaml

@@ -11,4 +11,4 @@ spec:
 ---
 kind: service.CodeCov
 spec:
-  targetThreshold: 15
+  targetThreshold: 30

api/v1alpha1/cluster.proto

@@ -18,6 +18,9 @@ service Cluster {
     // List affiliates in the cluster.
     rpc List(ListRequest) returns (ListResponse);
     // Watch affiliate updates in the cluster.
+    //
+    // The first WatchResponse contains the snapshot of the current affiliate state (even if empty).
+    // Subsequent response messages are sent as the affiliates get changed.
     rpc Watch(WatchRequest) returns (stream WatchResponse);
 }
 
@@ -108,7 +111,7 @@ message WatchRequest {
 
 message WatchResponse {
     // List of cluster affiliates.
-    Affiliate affiliate = 1;
+    repeated Affiliate affiliates = 1;
     // Flag that affiliate was deleted, only ID field is valid.
     bool deleted = 2;
 }