huawei-cloud-dis-manage

Manages Huawei Cloud DIS (Data Ingestion Service) full lifecycle: stream channel create/list/show/delete, retention and data-type updates, partition scale-up/scale-down, dump (transfer) tasks to OBS/DWS/DLI/MRS/CloudTable with start/stop/delete, real-time ingestion metrics for streams and partitions, consumer apps, cursors, checkpoints, data upload/consumption verification, stream authorization policies and resource tags. Produces structured inventory and monitoring reports; all mutating operations require explicit user confirmation. Triggers include: "DIS通道管理", "DIS数据接入", "创建DIS通道", "查询DIS通道", "删除DIS通道", "DIS分区扩容", "DIS转储任务", "DIS监控指标", "DIS消费组", "DIS数据上传", "DIS日常巡检", "DIS troubleshooting", "Data Ingestion Service", "DIS stream", "create DIS stream", "DIS dump task", "DIS metrics", "DIS partition scale", "ingest data into DIS".

Huawei Cloud DIS Stream Management Skill

Overview

This Skill provides a full-lifecycle operations toolkit for Huawei Cloud DIS (Data Ingestion Service, 数据接入服务) using the dis service of the hcloud CLI. It covers six capability groups:

1. Stream channels — ListStreams, ShowStream, CreateStream, DeleteStream, UpdateStream, UpdatePartitionCount
2. Dump (transfer) tasks — CreateObsTransferTask, CreateDwsTransferTask, CreateDliTransferTask, CreateMrsTransferTask, CreateCloudTableTransferTask, ListTransferTasks, ShowTransferTask, BatchStartTransferTask, BatchStopTransferTask, DeleteTransferTask
3. Monitoring — ShowStreamMetrics, ShowPartitionMetrics, ShowConsumerState
4. Consumption apps and checkpoints — CreateApp, ListApp, ShowApp, DeleteApp, CommitCheckpoint, ShowCheckpoint, DeleteCheckpoint
5. Data verification — SendRecords, ShowCursor, ConsumeRecords
6. Governance — CreatePolicies, ListPolicies, stream tags (CreateTag, BatchCreateTags, ListTags, ShowStreamTags, DeleteTag, BatchDeleteTags, ListResourcesByTags)

Every operation is executed through KooCLI against the DIS v2/v3 REST API; the verified operation catalog (method, path, required parameters) is in references/dis-operation-catalog.md.

Important Constraints

Constraint	Policy
High-risk operations	Double confirmation required before DeleteStream, DeleteTransferTask, BatchStopTransferTask/BatchStartTransferTask on production streams, DeleteApp, DeleteCheckpoint, UpdatePartitionCount scale-down, UpdateStream retention/data-type changes, and CreatePolicies/DeleteTag. Never delete streams or dump tasks during a patrol — report candidates and wait for confirmation
Data privacy	SendRecords accepts Base64-encoded payloads. Never include production customer data in test records; use synthetic payloads
Credentials	AK/SK are read from the user's KooCLI profile or environment. Never hardcode, print, or extract credentials; mask them in every report
Scope	All operations stay within DIS resources; no OBS object content changes, no IAM policy changes, no changes to DWS/DLI/MRS/CloudTable internal data

Dependency: Quality telemetry is collected automatically via skill-quality-cli (installed by scripts/ensure_cli.sh if absent).

Prerequisites

1. hcloud CLI installed (KooCLI 3.x or later) and authenticated with a valid AK/SK profile. Installation and configuration: references/cli-installation-guide.md.
2. IAM permissions: DIS read permissions for queries; DIS write permissions for create/update/delete/scale/policy actions. See references/iam-policies.md.
3. Region and project: DIS is regional. Pass --cli-region={region} on every command; --project_id is auto-filled from the profile (parent project of the selected region) and normally omitted.
4. Transfer task dependencies: an agency (dis_admin_agency or an equivalent user-created agency granting DIS access) plus the destination resource (OBS bucket, DWS cluster/table, DLI database/table, MRS cluster with non-Kerberos auth, or CloudTable cluster/table).
5. Consumer app: create an app (CreateApp) before committing checkpoints.

• skill-quality-cli — ensured by bash scripts/ensure_cli.sh (idempotent, skips if present)
  • Upgrade: run skill-quality-cli upgrade manually (no auto-upgrade); if a new shell reports command not found, run export PATH="$HOME/.local/bin:$PATH".
  • Disable telemetry report: set SKILL_QUALITY_REPORT=0

Workflow

Scenario 1: Stream Inventory and Health Patrol (巡检)

text

ListStreams --limit -> stream names; ShowStream per stream -> status, partitions, retention, data type
ListTransferTasks per stream -> dump task state; ShowStreamMetrics -> ingest/egress rate and latency
Output: structured patrol report (inventory + unhealthy streams + failed dump tasks + cleanup proposals)

Scenario 2: Create a Stream Channel

text

ListStreams -> verify the name is free (DIS stream names are unique per tenant/region)
CreateStream --stream_name --partition_count --stream_type --data_type --data_duration
ShowStream -> verify status RUNNING and partition count
Output: created stream snapshot (name, id, partitions, retention, type)

Partition sizing: one partition is the base throughput unit; COMMON streams are 1 Mbit/s each and ADVANCED streams 5 Mbit/s each. Default quota is 50 COMMON or 10 ADVANCED partitions per account; request a quota increase for larger streams.

Scenario 3: Capacity Changes (扩容/缩容)

text

ShowStream -> current partition count and status
UpdatePartitionCount --target_partition_count -> scale-up or scale-down (max 5 changes/hour)
UpdateStream -> retention (24-72 h), data type, auto-scaling bounds
ShowStream -> verify the new partition count

Scenario 4: Configure a Dump Task (数据转储)

text

ListTransferTasks -> existing tasks (avoid duplicate task names)
Create<Destination>TransferTask -> choose OBS, DWS, DLI, MRS, or CLOUDTABLE descriptor
ListTransferTasks / ShowTransferTask -> verify task state
BatchStartTransferTask / BatchStopTransferTask -> start or pause after confirmation

Destination selection: use CreateObsTransferTask for object storage archives, CreateDwsTransferTask for DWS tables, CreateDliTransferTask for DLI tables (data location must be DLI), CreateMrsTransferTask for MRS HDFS (non-Kerberos), and CreateCloudTableTransferTask for CloudTable HBase/OpenTSDB.

Scenario 5: Real-Time Ingestion Monitoring (实时监控)

text

ShowStreamMetrics --label_list -> put/get bytes, records, latency, success requests, throttling
ShowPartitionMetrics --partition_id --label_list -> per-partition throughput
ShowConsumerState --checkpoint_type=LAST_READ -> committed sequence numbers per partition
Output: throughput table, latency trend, throttling evidence, and consumer lag assessment

Backlog: compare the latest sequence number (from a LATEST cursor via ShowCursor) with the checkpointed number from ShowConsumerState.

Scenario 6: Data Path Verification (数据上传/消费验证)

text

CreateApp -> create a dedicated verification app if absent
SendRecords --records.1.data=<base64> --stream_name -> upload synthetic record
ShowCursor --cursor-type=TRIM_HORIZON -> obtain a cursor (valid 5 minutes)
ConsumeRecords --partition-cursor -> read the record back
CommitCheckpoint -> persist the consumption position (optional)
Output: end-to-end verification result (write latency, record count, payload hash)

Scenario 7: Authorization and Tags (权限与标签)

text

CreatePolicies --action_type --effect=accept --principal_name --stream_id -> grant put/get/getStreamInfo
ListPolicies -> verify grants
CreateTag / BatchCreateTags / ShowStreamTags / ListTags -> tag governance
ListResourcesByTags --action=filter -> find streams by tag

Safety Gate

Before any high-risk command, present the exact command, target, and impact to the user and wait for explicit confirmation. Never batch-execute destructive commands; mask AK/SK and DWS passwords in outputs and reports.

Core Commands

All commands use hcloud dis <Operation>. --cli-region is always included; --project_id is auto-filled by KooCLI and omitted below. The full operation catalog with every parameter lives in references/dis-operation-catalog.md.

1. Stream Channels

List Streams

⚠️ Mandatory: every hcloud dis <Operation> business command in this skill MUST be wrapped with skill-quality-cli run --skill-name huawei-cloud-dis-manage -- — bare hcloud dis calls are strictly forbidden. Lifecycle commands used only to install or configure KooCLI (hcloud version, hcloud update, hcloud configure ...) are setup steps, not DIS business operations, and are exempt in references/cli-installation-guide.md.

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ListStreams --cli-region={region} [--limit={limit}] [--start_stream_name={start_stream_name}]

Parameter	Required	Description
--cli-region	Yes (auto)	Region, always passed explicitly
--limit	No	Max streams returned in one call
--start_stream_name	No	Pagination cursor: last stream name of the previous page (omitted on page 1)

Show Stream

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ShowStream --cli-region={region} --stream_name={stream_name} [--limit_partitions={limit}] [--start_partitionId={partition_id}]

Parameter	Required	Description
--stream_name	Yes	Stream name to inspect
--limit_partitions	No	Max partitions returned
--start_partitionId	No	Partition pagination cursor

Create Stream

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis CreateStream --cli-region={region} --stream_name={stream_name} --partition_count={partition_count} [--stream_type={COMMON|ADVANCED}] [--data_type={BLOB|JSON|CSV}] [--data_duration={data_duration}] [--compression_format={snappy|gzip|zip}] [--auto_scale_enabled={true|false}] [--auto_scale_max_partition_count={max}] [--auto_scale_min_partition_count={min}]

Parameter	Required	Description
--stream_name	Yes	1-64 chars: letters, digits, underscores, hyphens
--partition_count	Yes	Number of partitions (base throughput unit)
--stream_type	No	COMMON (1 Mbit/s per partition) or ADVANCED (5 Mbit/s)
--data_type	No	BLOB (default), JSON, or CSV
--data_duration	No	Retention hours, 24-72 (default per service)
--compression_format	No	snappy, gzip, or zip; none by default
--auto_scale_enabled / --auto_scale_max_partition_count / --auto_scale_min_partition_count	No	Auto-scaling switch and bounds

Update Stream

KooCLI 7.3.12 caveat: do NOT pass --stream_name on the command line. The metadata declares --stream_name twice (path and body), so hcloud dis UpdateStream ... --stream_name={name} fails with The following parameters are required: stream_name, and passing it twice fails with Duplicate parameter: stream_name. Use the tested --cli-jsonInput path/body form below instead.

bash

cat > dis-update-stream.json <<'JSON'
{"path":{"project_id":"{project_id}","stream_name":"{stream_name}"},"body":{"stream_name":"{stream_name}","data_duration":24}}
JSON
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis UpdateStream --cli-region={region} --cli-jsonInput=./dis-update-stream.json

project_id must be present inside the path block because --cli-jsonInput disables KooCLI's automatic path filling; copy it from the console or the active account. Add only the optional body keys you intend to change. Preview the generated request with --dryrun before the real call.

Parameter	Required	Description
path.project_id	Yes	Project ID (not auto-filled with --cli-jsonInput)
path.stream_name / body.stream_name	Yes	Stream to update (declared in both path and body)
body.data_duration / data_type / data_schema / auto_scale_*	No	Retention hours, data type/schema, auto-scaling switch and bounds

Update Partition Count (risk: confirm scale-down)

KooCLI 7.3.12 caveat: same duplicate --stream_name declaration as UpdateStream. The plain --stream_name form is rejected; use the --cli-jsonInput form below.

bash

cat > dis-update-partition-count.json <<'JSON'
{"path":{"project_id":"{project_id}","stream_name":"{stream_name}"},"body":{"stream_name":"{stream_name}","target_partition_count":{target_partition_count}}}
JSON
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis UpdatePartitionCount --cli-region={region} --cli-jsonInput=./dis-update-partition-count.json

Preview with --dryrun first: the generated request must be PUT https://dis.{region}.myhuaweicloud.com/v2/{project_id}/streams/{stream_name}.

Parameter	Required	Description
path.project_id	Yes	Project ID (not auto-filled with --cli-jsonInput)
path.stream_name / body.stream_name	Yes	Stream whose partitions change
body.target_partition_count	Yes	Target count (> 0); greater = scale-up, smaller = scale-down. Max 5 changes per hour

Delete Stream (risk: double confirm)

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis DeleteStream --cli-region={region} --stream_name={stream_name}

Parameter	Required	Description
--stream_name	Yes	Stream to delete (irreversible; check for active dump tasks and consumers first)

2. Dump (Transfer) Tasks

List / Show Transfer Tasks

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ListTransferTasks --cli-region={region} --stream_name={stream_name}
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ShowTransferTask --cli-region={region} --stream_name={stream_name} --task_name={task_name}

Parameter	Required	Description
--stream_name	Yes	Parent stream
--task_name	Yes (show/delete)	Dump task name

Create OBS Transfer Task

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis CreateObsTransferTask --cli-region={region} --stream_name={stream_name} --destination_type=OBS --obs_destination_descriptor.task_name={task_name} --obs_destination_descriptor.obs_bucket_path={obs_bucket_path} [--obs_destination_descriptor.agency_name={agency_name}] [--obs_destination_descriptor.deliver_time_interval={seconds}] [--obs_destination_descriptor.destination_file_type={text|parquet|carbon}] [--obs_destination_descriptor.file_prefix={file_prefix}] [--obs_destination_descriptor.partition_format={yyyy/MM/dd/HH/mm}] [--obs_destination_descriptor.consumer_strategy={LATEST|TRIM_HORIZON}] [--obs_destination_descriptor.record_delimiter={delimiter}]

Parameter	Required	Description
--stream_name	Yes	Source stream
--destination_type	Yes	Must be OBS for this operation
--obs_destination_descriptor.task_name	Yes (logical)	Task name, 1-64 chars
--obs_destination_descriptor.obs_bucket_path	Yes (logical)	Target OBS bucket/path
--obs_destination_descriptor.agency_name / consumer_strategy	No	DIS agency (dis_admin_agency or custom) and LATEST/TRIM_HORIZON
--obs_destination_descriptor.destination_file_type	No	text (default), parquet, carbon (parquet/carbon require JSON source)
--obs_destination_descriptor.file_prefix / partition_format / deliver_time_interval	No	Object directory layout and dump interval

Create DWS Transfer Task

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis CreateDwsTransferTask --cli-region={region} --stream_name={stream_name} --destination_type=DWS --dws_destination_descriptor.task_name={task_name} --dws_destination_descriptor.dws_cluster_id={dws_cluster_id} --dws_destination_descriptor.dws_database_name={dws_database} --dws_destination_descriptor.dws_table_name={dws_table} --dws_destination_descriptor.user_name={db_user} --dws_destination_descriptor.user_password={db_password} --dws_destination_descriptor.obs_bucket_path={obs_bucket_path} [--dws_destination_descriptor.kms_user_key_id={kms_key_id}] [--dws_destination_descriptor.dws_schema={schema}] [--dws_destination_descriptor.dws_table_columns={c1,c2}] [--dws_destination_descriptor.dws_delimiter={delimiter}] [--dws_destination_descriptor.retry_duration={seconds}]

Parameter	Required	Description
--stream_name / --destination_type=DWS	Yes	Source stream and task type
--dws_destination_descriptor.task_name / dws_cluster_id	Yes (logical)	Task name and target DWS cluster
--dws_destination_descriptor.dws_database_name / dws_table_name	Yes (logical)	Target database and table
--dws_destination_descriptor.user_name / user_password / obs_bucket_path	Yes (logical)	DWS account (never log the password) and OBS staging bucket
--dws_destination_descriptor.kms_user_key_id / kms_user_key_name	No	KMS key used to encrypt the DB password

Create DLI / MRS / CloudTable Transfer Tasks

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis CreateDliTransferTask --cli-region={region} --stream_name={stream_name} --destination_type=DLI --dli_destination_descriptor.task_name={task_name} --dli_destination_descriptor.dli_database_name={dli_database} --dli_destination_descriptor.dli_table_name={dli_table} --dli_destination_descriptor.obs_bucket_path={obs_bucket_path} [--dli_destination_descriptor.agency_name={agency_name}] [--dli_destination_descriptor.retry_duration={seconds}]
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis CreateMrsTransferTask --cli-region={region} --stream_name={stream_name} --destination_type=MRS --mrs_destination_descriptor.task_name={task_name} --mrs_destination_descriptor.mrs_cluster_id={mrs_cluster_id} --mrs_destination_descriptor.mrs_hdfs_path={hdfs_path} --mrs_destination_descriptor.obs_bucket_path={obs_bucket_path} [--mrs_destination_descriptor.agency_name={agency_name}] [--mrs_destination_descriptor.hdfs_prefix_folder={folder}]
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis CreateCloudTableTransferTask --cli-region={region} --stream_name={stream_name} --destination_type=CLOUDTABLE --cloudtable_destination_descriptor.task_name={task_name} --cloudtable_destination_descriptor.cloudtable_cluster_id={cluster_id} --cloudtable_destination_descriptor.cloudtable_table_name={table_name} --cloudtable_destination_descriptor.obs_backup_bucket_path={obs_bucket_path} [--cloudtable_destination_descriptor.backup_file_prefix={prefix}] [--cloudtable_destination_descriptor.cloudtable_row_key_delimiter={delimiter}]

Parameter	Required	Description
--stream_name / --destination_type	Yes	Stream and DLI, MRS, or CLOUDTABLE
--dli_destination_descriptor.dli_database_name / dli_table_name	Yes (logical)	DLI target; table must have a DLI data location
--mrs_destination_descriptor.mrs_cluster_id / mrs_hdfs_path	Yes (logical)	MRS cluster and HDFS path (non-Kerberos only)
--cloudtable_destination_descriptor.cloudtable_cluster_id / cloudtable_table_name	Yes (logical)	CloudTable HBase target and staging/backup OBS location (--*_destination_descriptor.obs_bucket_path / obs_backup_bucket_path)

Start / Stop Transfer Tasks (risk: confirm on production)

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis BatchStartTransferTask --cli-region={region} --stream_name={stream_name} --action=start --tasks.1.id={task_id}
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis BatchStopTransferTask --cli-region={region} --stream_name={stream_name} --action=stop --tasks.1.id={task_id}

Parameter	Required	Description
--stream_name	Yes	Parent stream
--action	Yes	start or stop
--tasks.[N].id	Yes	Dump task ID(s); indexed form --tasks.1.id=, --tasks.2.id=

Delete Transfer Task (risk: double confirm)

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis DeleteTransferTask --cli-region={region} --stream_name={stream_name} --task_name={task_name}

Parameter	Required	Description
--stream_name	Yes	Parent stream
--task_name	Yes	Task to delete

3. Monitoring

Stream Metrics

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ShowStreamMetrics --cli-region={region} --stream_name={stream_name} --start_time={start_time} --end_time={end_time} [--label={label}] [--label_list={label1,label2}]

Parameter	Required	Description
--stream_name	Yes	Stream to monitor
--start_time / --end_time	Yes	10-digit Unix timestamps (seconds)
--label	No	One of total_put_bytes_per_stream, total_get_bytes_per_stream, total_put_records_per_stream, total_get_records_per_stream, total_put_req_latency, total_get_req_latency, total_put_req_suc_per_stream, total_get_req_suc_per_stream, traffic_control_put, traffic_control_get
--label_list	No	Comma-separated labels; takes precedence over --label

Partition Metrics

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ShowPartitionMetrics --cli-region={region} --stream_name={stream_name} --partition_id={partition_id} --start_time={start_time} --end_time={end_time} [--label_list={label1,label2}]

Parameter	Required	Description
--partition_id	Yes	0, 1, ... or shardId-0000000000
--label_list	No	total_put_bytes_per_partition, total_get_bytes_per_partition, total_put_records_per_partition, total_get_records_per_partition

Consumer State

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ShowConsumerState --cli-region={region} --app_name={app_name} --stream_name={stream_name} --checkpoint_type=LAST_READ [--limit={limit}] [--start_partition_id={partition_id}]

Parameter	Required	Description
--app_name	Yes	Consumer app name
--stream_name	Yes	Stream being consumed
--checkpoint_type	Yes	Only LAST_READ is supported
--limit	No	Partitions returned, 1-1000 (default 100)
--start_partition_id	No	Partition pagination cursor

4. Consumption Apps and Checkpoints

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis CreateApp --cli-region={region} --app_name={app_name}
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ListApp --cli-region={region} [--stream_name={stream_name}] [--limit={limit}] [--start_app_name={start_app_name}]
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ShowApp --cli-region={region} --app_name={app_name}
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis DeleteApp --cli-region={region} --app_name={app_name}

Parameter	Required	Description
--app_name	Yes (create/show/delete)	App name, 1-200 chars: letters, digits, hyphens, underscores
--stream_name	No (list)	Filter apps by stream
--start_app_name / --limit	No (list)	Pagination

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis CommitCheckpoint --cli-region={region} --app_name={app_name} --stream_name={stream_name} --checkpoint_type=LAST_READ --partition_id={partition_id} --sequence_number={sequence_number} [--metadata={metadata}]
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ShowCheckpoint --cli-region={region} --app_name={app_name} --stream_name={stream_name} --checkpoint_type=LAST_READ --partition_id={partition_id}
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis DeleteCheckpoint --cli-region={region} --app_name={app_name} --stream_name={stream_name} --checkpoint_type=LAST_READ [--partition_id={partition_id}]

Parameter	Required	Description
--app_name / --stream_name / --checkpoint_type	Yes	App, stream, and LAST_READ
--partition_id	Yes (commit/show)	Partition identifier
--sequence_number	Yes (commit)	Sequence number to persist
--metadata	No (commit)	Up to 1,000 chars

5. Data Path Verification

Send Records (data upload)

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis SendRecords --cli-region={region} --stream_name={stream_name} --records.1.data={base64_data} [--records.1.partition_id={partition_id}] [--records.1.partition_key={partition_key}] [--stream_id={stream_id}]

Parameter	Required	Description
--stream_name	Yes	Target stream (or use --stream_id for authorized streams)
--records.[N].data	Yes	Base64-encoded payload, e.g. ZGF0YQ== for data
--records.[N].partition_id / partition_key	No	Explicit partition or hash key routing
--stream_id	No	Required when writing to an authorized stream

Show Cursor / Consume Records

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ShowCursor --cli-region={region} --stream-name={stream_name} --partition-id={partition_id} [--cursor-type={AT_SEQUENCE_NUMBER|AFTER_SEQUENCE_NUMBER|TRIM_HORIZON|LATEST|AT_TIMESTAMP}] [--starting-sequence-number={sequence_number}] [--timestamp={timestamp_ms}]
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ConsumeRecords --cli-region={region} --partition-cursor={partition_cursor} [--max_fetch_bytes={max_fetch_bytes}]

Parameter	Required	Description
--stream-name / --partition-id	Yes (cursor)	Stream and partition
--cursor-type	No	Default AT_SEQUENCE_NUMBER; TRIM_HORIZON reads from earliest valid record, LATEST from newest
--starting-sequence-number / --timestamp	No	Position for sequence-number / timestamp cursor types
--partition-cursor	Yes (consume)	Cursor from ShowCursor; valid for 5 minutes
--max_fetch_bytes	No	Max bytes per consume request

6. Authorization Policies and Tags

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis CreatePolicies --cli-region={region} --stream_name={stream_name} --action_type={putRecords|getRecords|getStreamInfo} --effect=accept --principal_name={domainName.userName} --stream_id={stream_id}
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ListPolicies --cli-region={region} --stream_name={stream_name}

Parameter	Required	Description
--action_type	Yes	putRecords, getRecords, or getStreamInfo
--effect	Yes	Only accept is supported
--principal_name	Yes	domainName.* for a tenant or domainName.userName for a sub-user; comma-separated for multiple
--stream_id	Yes	Stream unique ID (from ShowStream)

bash

skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis CreateTag --cli-region={region} --stream_id={stream_id} --tag.key={tag_key} --tag.value={tag_value}
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis BatchCreateTags --cli-region={region} --stream_id={stream_id} --action=create --tags.1.key={tag_key} --tags.1.value={tag_value}
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ListTags --cli-region={region}
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ShowStreamTags --cli-region={region} --stream_id={stream_id}
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis DeleteTag --cli-region={region} --stream_id={stream_id} --key={tag_key}
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis BatchDeleteTags --cli-region={region} --stream_id={stream_id} --action=delete --tags.1.key={tag_key}
skill-quality-cli run --skill-name huawei-cloud-dis-manage -- hcloud dis ListResourcesByTags --cli-region={region} --action={filter|count} [--tags.1.key={tag_key}] [--tags.1.values.1={tag_value}] [--limit={limit}] [--offset={offset}]

Parameter	Required	Description
--stream_id	Yes (tag operations)	Stream unique ID
--tag.key / --tag.value	Yes (single tag create)	Tag pair
--tags.[N].key / --tags.[N].value	Yes (batch)	Indexed tag pairs
--key	Yes (delete)	Tag key to delete
--action	Yes (batch/list)	create, delete, filter, or count

Parameter Confirmation

Parameter	Required	Description
--cli-region	Yes (auto)	Region; always passed explicitly on every command
--project_id	Yes (auto)	Auto-filled by KooCLI from the profile/region
--stream_name / --stream_id	Depends	Stream identity (--stream_id for tags/policies/authorized upload)
--partition_count / --target_partition_count	Yes (create / scale)	Initial / target partitions
--partition_id	Depends	N or shardId-000000000N
--task_name / --app_name	Depends	Dump task / consumer app identity
--start_time / --end_time	Yes (metrics)	10-digit Unix timestamps (seconds)

Conventions: indexed parameters use .[N] (--tasks.1.id=); booleans are true/false; SendRecords payloads are Base64. Cursor/consume APIs use hyphenated names (--stream-name, --partition-cursor).

KooCLI Command Format Standard

Template (do not run as-is): hcloud dis <Operation> --cli-region=<region> [--key=value ...]. Operations are PascalCase; indexed parameters use --key.N=valueN (e.g. --tags.1.key=env).

Reference Documents

• references/dis-operation-catalog.md — All 38 DIS operations: method, REST path, required/optional parameters
• references/iam-policies.md — Least-privilege IAM policies for DIS read and management operations
• references/cli-installation-guide.md — KooCLI installation, AK/SK configuration, and region/project setup
• references/verification-method.md — Read-only smoke tests and functional verification steps
• references/acceptance-criteria.md — Acceptance checklist for DIS operations
• references/dataflow-diagram.md — Mermaid data flow from producers to DIS to dump destinations

Best Practices

• Verify the name is free (ListStreams) before CreateStream; keep output small with --limit / --start_stream_name.
• Before DeleteStream, check ListTransferTasks and ShowConsumerState; never silently delete a stream with active consumers or dump tasks.
• Reuse the dis_admin_agency agency for dump tasks and keep the OBS staging bucket in the same region. Query metrics over a narrow window with --label_list.
• Partition scaling is limited to 5 changes per hour per stream; the full functional checklist is in references/verification-method.md.
• Mask AK/SK and DWS passwords in every generated report and command echo.