alibabacloud-sls-dashboard

Design and manage Alibaba Cloud SLS dashboards and visualizations. Use when creating or editing dashboards and charts, explaining visualization settings, troubleshooting display issues, or managing scheduled dashboard reports.

SLS Dashboards

Design dashboards around the user's questions and verified SLS data. Resolve script paths relative to this file.

Prerequisites

Explanation, offline construction and validation do not require CLI setup or cloud credentials. Offline scripts still need the Python dependencies. Before the first build or cloud operation, run python3 scripts/doctor.py with the Python interpreter that will run the task. It checks Python dependencies, Aliyun CLI 3.3.22 or later, CLI profile configuration, the SLS plugin, and dashboard/query command help, and prints versions and suggested commands. Use --offline for offline work; add --with-sdk for Metric StoreView queries or scheduled reports. See environment setup for details.

For cloud operations, use the configured profile by default; add --profile when the user specifies one. The caller needs the RAM permissions for the operation.

Observability (required for cloud operations)

Before the first cloud call, read manifest.json and use its non-empty version as {skill-version}. Stop if the file or version is invalid; do not guess a version.

Generate one random 32-character lowercase hexadecimal {session-id} for this skill in the conversation. Reuse it for all operations, pagination and retries. When switching skills, use each skill's own version and session ID, including on return.

Set ALIBABACLOUD_SLS_DASHBOARD_USERAGENT to:

text

AlibabaCloud-Agent-Skills/alibabacloud-sls-dashboard/{session-id} skill-version/{skill-version}

The bundled scripts use this variable for cloud requests. For direct cloud API commands, including request-only dry runs, add:

sh

--user-agent "$ALIBABACLOUD_SLS_DASHBOARD_USERAGENT"

Local help, version, plugin and configure commands, offline dashboard operations, and doctor checks do not need it. Never send literal placeholders, credentials or personal identifiers.

Choose a workflow

Request	Read
Explain settings or review JSON/Plan	Consult
Create a dashboard or build a supplied Plan	Create
Edit, add, remove charts, or combine changes	Edit
Publish or delete a dashboard	Publish
Diagnose missing data or broken controls	Diagnose
Manage scheduled reports	Subscriptions

Use the knowledge index to select references. Explain each setting's purpose and effect. Include commands or configuration fields when needed to complete the task; omit implementation details.

For a simple starting configuration, read common chart examples and open the selected single-chart grid Dashboard JSON. Follow its links for detailed settings.

For CLI or script failures, use skill usage troubleshooting.

Working guidelines

• Plan builds and online operations support SLS Logstore, Metricstore and existing named StoreViews. The capability table lists tool support and other Dashboard JSON sources, including builtin simulated data.
• Keep the destination Project/Region independent of query sources. Preserve StoreView identity instead of replacing a view with a member store.
• Use resource names, fields, metrics and labels from user input, existing configuration or discovery. A sampled field does not prove SQL indexing.
• Prefer a Plan for new content. Complete Dashboard JSON can be validated directly. Edit existing JSON without regenerating unrelated content. Use lowercase stable IDs for new charts; see chart naming.
• Use grid for new dashboards; use free only when a custom layout is needed. Set attribute.type explicitly. Validation accepts empty existing layout settings as free; see layout.
• Display-only edits need no datasource discovery. Field changes must account for query results, transformations and display bindings. Controls need matching consumers.
• Keep an original baseline for edits and review the diff before publishing.
• Build and validation are offline. Publish/delete require user authorization and --execute; an explicit request to publish authorizes publishing. Creating a dashboard does not authorize report deliveries.
• Report the checks performed. Saving successfully does not prove that queries or charts render correctly.

Run the tools

Use Python 3.10+ with the Python dependencies installed. Metric StoreView queries and scheduled reports also need the optional packages in that guide. Offline work needs no cloud credentials.

• python3 scripts/sls.py --help: resources, discovery, queries and dashboard reads.
• python3 scripts/dashboard.py --help: build, edit checks, publishing and subscriptions.
• python3 scripts/validator.py --help: individual chart or dashboard configuration checks.

Cloud commands accept --profile and --endpoint after the operation name; --region remains required. See commands and connection options.