[Admin API] Add new rolling restart health endpoints

modules/ROOT/attachments/admin-api.yaml

@@ -775,6 +775,41 @@ paths:
         200:
           description: Recommission broker success
           content: {}
+  /v1/broker/pre_restart_probe:
+    get:
+      tags:
+        - Brokers
+      summary: Check broker pre-restart
+      description: Check if it is safe to restart this broker. 
+      operationId: pre_restart_probe
+      parameters:
+        - name: limit
+          in: query
+          required: false
+          schema:
+            type: integer
+          description: "Limit the number of partitions listed for each risk type (default: 128)."
+      responses:
+        200:
+          description: Pre-restart check result. Returns risks associated with restarting the broker, and partitions affected, if any.
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/pre_restart_check_result'
+  /v1/broker/post_restart_probe:
+    get:
+      tags:
+      - Brokers
+      summary: Check broker post-restart
+      description: Check if the broker has recovered after a restart.
+      operationId: post_restart_probe
+      responses:
+        200:
+          description: Post-restart check result
+          content:
+            application/json:
+              schema:
+                $ref: '#/components/schemas/post_restart_check_result'
   /v1/cluster_view:
     get:
       tags:
@@ -2249,6 +2284,48 @@ components:
           type: string
         port:
           type: integer
+    pre_restart_check_result:
+      type: object
+      description: Pre-restart check result.
+      properties:
+        risks:
+          $ref: '#/components/schemas/restart_risks'
+    restart_risks:
+      description: Partitions affected by the current broker restart, grouped by risk type. Each partition list is truncated according to the optional limit specified in the request.
+      properties:
+        rf1_offline:
+          type: array
+          items:
+            type: string
+            description: Namespace, topic, partition ID
+          description: Partitions with a replication factor of 1 that have a replica on the current broker.
+        full_acks_produce_unavailable:
+          type: array
+          items:
+            type: string
+            description: Namespace, topic, partition ID
+          description: Partitions that may reject produce requests (with `acks=-1`) if the current broker is restarted.
+        unavailable:
+          type: array
+          items:
+            type: string
+            description: Namespace, topic, partition ID
+          description: Partitions that may reject consume and produce requests if the current broker is restarted.
+        acks1_data_loss:
+          type: array
+          items:
+            type: string
+            description: Namespace, topic, partition ID
+          description: Partitions that may lose data produced (with `acks=1`) if the current broker is restarted. 
+    post_restart_check_result:
+      type: object
+      description: Post-restart check result.
+      properties:
+        load_reclaimed_pc:
+          type: integer
+          description: The load that the broker has reclaimed after restarting, as a percentage of in-sync replicas.
+          minimum: 0
+          maximum: 100
     broker_locator:
       type: object
       properties: