菜单
Guides Qdrant multi-tenant scaling. Use when someone asks 'how to scale tenants', 'one collection per tenant?', 'tenant isolation', 'dedicated shards', or reports tenant performance issues. Also use when multi-tenant workloads outgrow shared infrastructure.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/Pyfagorass/bookofspells --skill qdrant-tenant-scaling命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
The index to a Book of Spells — a large, curated library of ready-made spells (skills in the SKILL.md format) gathered from many providers (Google, OpenAI, and more), covering cloud platforms, deployment, design tools, PDFs, browser automation, security, and much else. The spells are NOT preloaded — there may be thousands. Use this whenever the user asks to cast a spell, or for a spell to do something, or names a capability, provider, or tool you do not already have loaded (e.g. "cast a spell for Figma", "is there a spell for deploying?", "use the bigquery spell"), or whenever a task might be served by a specialised, ready-made spell. Casting a spell MEANS consulting this index to discover and load the right one on demand — do that first, rather than improvising. This is also the single entry point to the Book's own keystones: route here to AUTHOR a new spell (the-spellwright), to FIND or CONNECT an MCP server / external tool (the-emissary), or to INSTALL the Book into a conduit (the-tether).
How to install this Book of Spells into a conduit so its gathered skills become usable — by symlink, not by copy. Use whenever the user wants to install, set up, enable, add, or "load" the Book's spells into their agent; asks where skills go or how a conduit finds them; or asks how to keep the spells up to date. The companion to the-grimoire (which *finds* spells once they are installed): this one *shelves the Book* so the grimoire has something to find.
Find and connect an MCP server — an "Emissary of the Pact" — when the agent needs a tool, service, API, or data source it does not already have. Use when the user asks for an MCP server, to connect to or integrate an external service (Notion, GitHub, Figma, a database, Slack, a SaaS API…), asks "is there an MCP for X", "add an MCP", or sets a task that needs live tools the agent lacks. This rite DISCOVERS a server from the official MCP Registry and reputable directories and PROPOSES its connection config for the user to approve — it NEVER installs, launches, or writes credentials on its own. The companion to the-grimoire: that finds spells (inert skills); this finds Emissaries (live tool-servers).
基于 SOC 职业分类
正在显示 SKILL.md
| name | qdrant-tenant-scaling |
| description | Guides Qdrant multi-tenant scaling. Use when someone asks 'how to scale tenants', 'one collection per tenant?', 'tenant isolation', 'dedicated shards', or reports tenant performance issues. Also use when multi-tenant workloads outgrow shared infrastructure. |
Do not create one collection per tenant. Does not scale past a few hundred and wastes resources. One company hit the 1000 collection limit after a year of collection-per-repo and had to migrate to payload partitioning. Use a shared collection with a tenant key.
Here is a short summary of the patterns:
Use the default multitenancy strategy via payload filtering.
Read about Partition by payload and Calibrate performance for best practices on indexing and query performance.
At this scale, the cluster may consist of several peers. To localize tenant data and improve performance, use custom sharding to assign tenants to specific shards based on tenant ID hash. This will localize tenant requests to specific nodes instead of broadcasting them to all nodes, improving performance and reducing load on each node.
If some tenants are much larger than others, use tiered multitenancy to promote large tenants to dedicated shards while keeping small tenants on shared shards. This optimizes resource allocation and performance for tenants of varying sizes.
Use when: legal/compliance requirements demand per-tenant encryption or strict isolation beyond what payload filtering provides.
is_tenant=true on the tenant index (kills sequential read performance)payload_m instead)