Introducing pipeline uid in indexing tasks/ indexing pipelines.

Introducing pipeline uid in indexing tasks/ indexing pipelines.

The scheduling algorithm is for the most part unchanged.
Only the allocation of shard to pipelines is modified.

Currently the logic yields similar results as before. In particular, the scheduling algorithm
may in some case, move shard from one pipeline to another.
In the future, we will prevent this behavior and instead close/open new shards.

Closes #4191

quickwit/quickwit-cli/src/tool.rs

@@ -52,7 +52,7 @@ use quickwit_metastore::IndexMetadataResponseExt;
 use quickwit_proto::indexing::CpuCapacity;
 use quickwit_proto::metastore::{IndexMetadataRequest, MetastoreService, MetastoreServiceClient};
 use quickwit_proto::search::{CountHits, SearchResponse};
-use quickwit_proto::types::NodeId;
+use quickwit_proto::types::{NodeId, PipelineUid};
 use quickwit_search::{single_node_search, SearchResponseRest};
 use quickwit_serve::{
     search_request_from_api_request, BodyFormat, SearchRequestQueryString, SortBy,
@@ -473,7 +473,7 @@ pub async fn local_ingest_docs_cli(args: LocalIngestDocsArgs) -> anyhow::Result<
         .ask_for_res(SpawnPipeline {
             index_id: args.index_id.clone(),
             source_config,
-            pipeline_ord: 0,
+            pipeline_uid: PipelineUid::from_u128(0u128),
         })
         .await?;
     let merge_pipeline_handle = indexing_server_mailbox
@@ -613,7 +613,7 @@ pub async fn merge_cli(args: MergeArgs) -> anyhow::Result<()> {
                 transform_config: None,
                 input_format: SourceInputFormat::Json,
             },
-            pipeline_ord: 0,
+            pipeline_uid: PipelineUid::from_u128(0u128),
         })
         .await?;
     let pipeline_handle: ActorHandle<MergePipeline> = indexing_service_mailbox