Storing necessary fields instead of the entire QueryContext struct

[WenyXu]

          It's better to store only the necessary fields instead of the entire `QueryContext` struct. I'm going to convert this PR into an issue and plan to refactor it in the future.

Originally posted by @WenyXu in #5200 (comment)

[evenyag]

@WenyXu Can we turn this into a good first issue if you provide more context for it?

[WenyXu]

greptimedb/src/common/meta/src/ddl/create_flow.rs

Lines 326 to 338 in 52eebfc

	#[derive(Debug, Serialize, Deserialize)]
	pub struct CreateFlowData {
	pub(crate) cluster_id: ClusterId,
	pub(crate) state: CreateFlowState,
	pub(crate) task: CreateFlowTask,
	pub(crate) flow_id: Option<FlowId>,
	pub(crate) peers: Vec<Peer>,
	pub(crate) source_table_ids: Vec<TableId>,
	pub(crate) query_context: QueryContext,
	/// For verify if prev value is consistent when need to update flow metadata.
	/// only set when `or_replace` is true.
	pub(crate) prev_flow_info_value: Option<DeserializedValueWithBytes<FlowInfoValue>>,
	}

We don't need to store the entire QueryContext, only store the required fields of QueryContext in CreateFlowData. e.g., schema, catalog, timezone.

greptimedb/src/common/meta/src/rpc/ddl.rs

Lines 1205 to 1212 in 52eebfc

	#[derive(Debug, Clone, Serialize, Deserialize)]
	pub struct QueryContext {
	current_catalog: String,
	current_schema: String,
	timezone: String,
	extensions: HashMap<String, String>,
	channel: u8,
	}