菜单
This skill should be used when the user asks about "Temporal sizing", "history shards", "cluster capacity", "Temporal resources", "scale Temporal", "Temporal performance", "how many shards", or needs guidance on capacity planning for Temporal clusters.
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。
基于 SOC 职业分类
This skill should be used when the user asks about deploying Hugo to AWS S3, S3 static site hosting, Hugo with CloudFront, the hugo deploy command, configuring Hugo's deployment section, GitHub Actions with AWS OIDC for Hugo, S3 bucket static website configuration, CloudFront cache invalidation after Hugo deploy, or choosing between GitHub Pages and S3 for Hugo hosting. Also applies when GitHub Pages limitations (auth, custom headers, size) require an alternative deployment target.
This skill should be used when the user asks about deploying Hugo to GitHub Pages, setting up a GitHub Actions workflow for Hugo, Hugo CI/CD pipeline, content-aware path filtering for Hugo builds, caching Hugo modules in CI, Hugo PR preview deployments, Hugo build workflow, or configuring GitHub Pages deployment for a Hugo site. Also applies when troubleshooting GitHub Actions workflow failures related to Hugo builds.
This skill should be used when the user asks about Hugo's data directory, loading YAML or JSON or TOML data files in Hugo templates, using .Site.Data to access structured data, data-driven navigation or menus, rendering datasets or bibliographies or registries from data files, generating content from data, using range to iterate over data collections, Hugo resources.GetRemote for external data, or organizing data files for large projects. Also applies when the user wants to drive site content or navigation from structured data rather than hardcoding it in templates.
This skill should be used when the user asks about Hugo shortcodes, creating custom shortcodes, Hugo render hooks, custom rendering for links or images or headings or code blocks, Hugo taxonomies (tags, categories, custom), page bundles and page resources, content formatting beyond basic markdown, creating callout boxes or tabs or admonitions in Hugo, or customizing how Hugo processes markdown elements. Also applies when the user wants to add reusable content components to their Hugo site.
This skill should be used when the user asks about Hugo themes, installing a theme, customizing a theme, overriding theme layouts or partials, Hugo template lookup order, creating a Hugo theme from scratch, Hugo asset pipeline (SCSS, PostCSS), Hugo baseof.html, or theme configuration parameters. Also applies when evaluating which theme to use, switching themes, or troubleshooting theme-related rendering issues.
This skill should be used when the user asks about Hugo module mounts, mounting multiple directories into Hugo's content tree, aggregating docs from scattered directories, using hugo mod init, configuring module.mounts in hugo.toml, building a unified documentation site from a monorepo, mapping subdirectory docs/ folders into a single Hugo site, or understanding Hugo's union file system. Also applies when the user has a repository with multiple docs/ directories that need to be presented as a single site. This is the core skill for multi-directory content aggregation.
| name | cluster-sizing |
| description | This skill should be used when the user asks about "Temporal sizing", "history shards", "cluster capacity", "Temporal resources", "scale Temporal", "Temporal performance", "how many shards", or needs guidance on capacity planning for Temporal clusters. |
| version | 1.0.0 |
Guidance for sizing Temporal clusters based on workload requirements.
| Factor | Impact | Cannot Change |
|---|---|---|
| History Shards | Workflow parallelism | Yes (set at creation) |
| History Replicas | Throughput, availability | No |
| Matching Replicas | Task dispatch rate | No |
| Frontend Replicas | API request rate | No |
| Database Size | History storage | No |
Critical: History shards cannot be changed after cluster creation.
Shards determine maximum workflow parallelism. Each workflow belongs to one shard.
| Concurrent Workflows | Recommended Shards |
|---|---|
| < 10,000 | 128 |
| 10,000 - 100,000 | 256 |
| 100,000 - 500,000 | 512 |
| 500,000 - 2,000,000 | 1024 |
| > 2,000,000 | 2048 or 4096 |
shards = ceil(max_concurrent_workflows / 1000) * safety_factor
# Round up to nearest power of 2
# safety_factor = 2-4x for growth
Example: Expecting 50,000 concurrent workflows with 3x growth:
base = 50,000 / 1000 = 50
with_growth = 50 * 3 = 150
nearest_power_of_2 = 256 shards
Shards distribute across history service replicas:
shards_per_replica = total_shards / history_replicas
# Example: 512 shards, 4 replicas = 128 shards/replica
More replicas = better distribution = higher throughput.
Handles API requests, authentication, rate limiting.
| Load Level | Replicas | CPU | Memory |
|---|---|---|---|
| Low (<100 rps) | 1-2 | 500m | 1Gi |
| Medium (100-1000 rps) | 3 | 1 | 2Gi |
| High (1000-5000 rps) | 5 | 2 | 4Gi |
| Very High (>5000 rps) | 10+ | 4 | 8Gi |
Manages workflow state and event history.
| Shards | Replicas | CPU/replica | Memory/replica |
|---|---|---|---|
| 128 | 2 | 1 | 2Gi |
| 256 | 3 | 2 | 4Gi |
| 512 | 4-6 | 2 | 4Gi |
| 1024 | 8-12 | 4 | 8Gi |
| 2048 | 16-24 | 4 | 8Gi |
Dispatches tasks to workers.
| Task Rate | Replicas | CPU | Memory |
|---|---|---|---|
| Low (<1000/s) | 2 | 500m | 1Gi |
| Medium (1000-10000/s) | 3 | 1 | 2Gi |
| High (>10000/s) | 5+ | 2 | 4Gi |
Handles internal system workflows. Scale with cluster size:
| Cluster Size | Replicas | CPU | Memory |
|---|---|---|---|
| Small | 1 | 200m | 256Mi |
| Medium | 1 | 500m | 512Mi |
| Large | 2 | 1 | 1Gi |
| Workflow Volume | CPU | Memory | Storage | IOPS |
|---|---|---|---|---|
| < 100K workflows | 2 | 8GB | 100GB | 3000 |
| 100K-1M workflows | 4 | 16GB | 500GB | 6000 |
| 1M-10M workflows | 8 | 32GB | 1TB | 12000 |
| > 10M workflows | 16+ | 64GB+ | 2TB+ | 20000+ |
storage_per_workflow = avg_history_events * event_size
= 100 events * 1KB = 100KB
total_storage = workflows * storage_per_workflow * retention_multiplier
= 1,000,000 * 100KB * 1.5 = 150GB
Retention: Configure appropriate workflow retention to manage storage.
For visibility queries (optional but recommended):
| Indexed Workflows | Nodes | CPU/node | Memory/node | Storage/node |
|---|---|---|---|---|
| < 1M | 3 | 1 | 2Gi | 50Gi |
| 1M-10M | 3 | 2 | 4Gi | 200Gi |
| > 10M | 5+ | 4 | 8Gi | 500Gi |
server:
config:
numHistoryShards: 128
replicaCount:
frontend: 1
history: 1
matching: 1
worker: 1
resources:
frontend:
requests: {cpu: "250m", memory: "512Mi"}
history:
requests: {cpu: "500m", memory: "1Gi"}
matching:
requests: {cpu: "250m", memory: "512Mi"}
server:
config:
numHistoryShards: 256
replicaCount:
frontend: 3
history: 3
matching: 3
worker: 1
resources:
frontend:
requests: {cpu: "500m", memory: "1Gi"}
limits: {cpu: "2", memory: "4Gi"}
history:
requests: {cpu: "1", memory: "2Gi"}
limits: {cpu: "4", memory: "8Gi"}
matching:
requests: {cpu: "500m", memory: "1Gi"}
limits: {cpu: "2", memory: "4Gi"}
server:
config:
numHistoryShards: 1024
replicaCount:
frontend: 5
history: 10
matching: 5
worker: 2
resources:
frontend:
requests: {cpu: "2", memory: "4Gi"}
limits: {cpu: "4", memory: "8Gi"}
history:
requests: {cpu: "4", memory: "8Gi"}
limits: {cpu: "8", memory: "16Gi"}
matching:
requests: {cpu: "2", memory: "4Gi"}
limits: {cpu: "4", memory: "8Gi"}
Scale replicas when:
Increase resources when:
Key metrics to watch:
# History service load
sum(rate(temporal_persistence_requests_total[5m])) by (operation)
# Task latency (indicates matching capacity)
histogram_quantile(0.99, rate(temporal_schedule_to_start_latency_bucket[5m]))
# Workflow throughput
sum(rate(temporal_workflow_completed_total[5m]))
# Shard distribution
temporal_history_shard_count
| Mistake | Impact | Solution |
|---|---|---|
| Too few shards | Cannot scale later | Start with more shards |
| Undersized history | Latency spikes | Increase memory, replicas |
| Single frontend | Single point of failure | Minimum 2 for HA |
| No Elasticsearch | Slow visibility queries | Enable for production |
For detailed sizing calculations, consult:
references/sizing-calculator.md - Detailed sizing formulasreferences/benchmark-results.md - Performance benchmark data