Fix cat allocation API response (#2847)

Author: pquentin

output/openapi/elasticsearch-openapi.json

@@ -17,6 +17,7 @@
  * under the License.
  */
 
+import { Stringified } from '@spec_utils/Stringified'
 import { ByteSize } from '@_types/common'
 import { Host, Ip } from '@_types/Networking'
 import { double, Percentage } from '@_types/Numeric'
@@ -26,73 +27,73 @@ export class AllocationRecord {
    * Number of primary and replica shards assigned to the node.
    * @aliases s
    */
-  shards: string
+  shards?: string
 
   /**
    * Amount of shards that are scheduled to be moved elsewhere in the cluster or -1 other than desired balance allocator is used
    */
-  'shards.undesired': string | null
+  'shards.undesired'?: string | null
 
   /**
    * Sum of index write load forecasts
    * @aliases wlf,writeLoadForecast
    */
-  'write_load.forecast': double | null
+  'write_load.forecast'?: Stringified<double> | null
 
   /**
    * Sum of shard size forecasts
    * @aliases dif,diskIndicesForecast
    */
-  'disk.indices.forecast': ByteSize | null
+  'disk.indices.forecast'?: ByteSize | null
   /**
    * Disk space used by the node’s shards. Does not include disk space for the translog or unassigned shards.
    * IMPORTANT: This metric double-counts disk space for hard-linked files, such as those created when shrinking, splitting, or cloning an index.
    * @aliases di,diskIndices
    */
-  'disk.indices': ByteSize | null
+  'disk.indices'?: ByteSize | null
   /**
    * Total disk space in use.
    * Elasticsearch retrieves this metric from the node’s operating system (OS).
    * The metric includes disk space for: Elasticsearch, including the translog and unassigned shards; the node’s operating system; any other applications or files on the node.
    * Unlike `disk.indices`, this metric does not double-count disk space for hard-linked files.
    * @aliases du,diskUsed
    */
-  'disk.used': ByteSize | null
+  'disk.used'?: ByteSize | null
   /**
    * Free disk space available to Elasticsearch.
    * Elasticsearch retrieves this metric from the node’s operating system.
    * Disk-based shard allocation uses this metric to assign shards to nodes based on available disk space.
    * @aliases da,diskAvail
    */
-  'disk.avail': ByteSize | null
+  'disk.avail'?: ByteSize | null
   /**
    * Total disk space for the node, including in-use and available space.
    * @aliases dt,diskTotal
    */
-  'disk.total': ByteSize | null
+  'disk.total'?: ByteSize | null
   /**
    * Total percentage of disk space in use. Calculated as `disk.used / disk.total`.
    * @aliases dp,diskPercent
    */
-  'disk.percent': Percentage | null
+  'disk.percent'?: Percentage | null
   /**
    * Network host for the node. Set using the `network.host` setting.
    * @aliases h
    */
-  host: Host | null
+  host?: Host | null
   /**
    * IP address and port for the node.
    */
-  ip: Ip | null
+  ip?: Ip | null
   /**
    * Name for the node. Set using the `node.name` setting.
    * @aliases n
    */
-  node: string
+  node?: string
 
   /**
    * Node roles
    * @aliases r,role,nodeRole
    */
-  'node.role': string | null
+  'node.role'?: string | null
 }