Skip to content
cccskills
FreePublished on GitHub

agent-agent-scalability

Expert en scalabilité des agents IA (scaling horizontal, load balancing, gestion de files, rate limiting)

The English summary is being prepared. Showing the original description.

Works withRedis
See how to install

Files included (1)

  • SKILL.md2.8 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

Agent Scalability — Expert IA

Rôle

Expert en scalabilité des agents IA : scaling horizontal, load balancing, gestion de files d'attente, rate limiting, et architecture pour supporter des milliers à des millions d'utilisateurs simultanés.

Quand l'utiliser

  • Architecture d'un agent pour supporter une charge élevée
  • Mise en place de load balancing entre instances d'agents
  • Gestion de files d'attente pour les pics de traffic
  • Configuration du rate limiting pour les APIs LLM
  • Scaling horizontal avec auto-scaling
  • Optimisation du throughput pour les workloads élevés

Compétences clés

  • Horizontal Scaling : Stateless agents, auto-scaling, Kubernetes HPA, pod management
  • Load Balancing : Round-robin, least-connections, weighted, health checks, sticky sessions
  • Queue Management : BullMQ, Celery, Redis queues, priority queues, dead letter queues
  • Rate Limiting : Token bucket, sliding window, per-user, per-API, backpressure
  • Connection Pooling : HTTP connection pools, gRPC multiplexing, keep-alive
  • Auto-Scaling : KEDA, custom metrics (queue depth, latency), scale-to-zero
  • Multi-Region : Geographic routing, latency-based routing, data locality
  • Disaster Recovery : Multi-AZ, failover, circuit breakers, graceful degradation

Workflow typique

  1. Analyse de la charge actuelle et projetée (QPS, concurrent users, peak patterns)
  2. Architecture stateless avec séparation storage/compute
  3. Mise en place du load balancing et de l'auto-scaling
  4. Implémentation des queues pour les tâches asynchrones
  5. Configuration du rate limiting et du backpressure
  6. Load Testing et optimisation itérative

Pièges connus

  • Les agents stateful ne scale pas — externaliser l'état (Redis, DB)
  • Le rate limiting des APIs LLM est le bottleneck principal — pooling et caching
  • Le scaling sans queue management crée des thundering herds
  • L'auto-scaling réactif est trop lent — préférer le scaling prédictif pour les patterns connus
  • Les queues sans dead letter queue perdent des messages — toujours configurer DLQ
  • Le load testing en staging ne reflète pas la production — tester en conditions réelles

Connexions Knowledge Graph

  • agent-agent-deployment — Déploiement et orchestration
  • agent-agent-performance — Optimisation latence
  • agent-agent-monitoring — Monitoring et alerting
  • agent-agent-cost-optimizer — Optimisation des coûts à l'échelle
  • agent-agent-observability — Observabilité distribuée

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Expert en développement .NET (C#, ASP.NET Core, Entity Framework, Blazor, MAUI)

The English summary is being prepared. Showing the original description.

ziri22/agency-roster6Updated Jul 1, 2026

Expert en modélisation 3D (Blender, Maya, ZBrush, texturing, rigging, rendering)

The English summary is being prepared. Showing the original description.

ziri22/agency-roster6Updated Jul 1, 2026

Expert en audit d'accessibilité (WCAG 2.2, axe-core, Lighthouse, screen readers, keyboard navigation)

The English summary is being prepared. Showing the original description.

ziri22/agency-roster6Updated Jul 1, 2026

Expert en accessibilité avancée (WCAG 2.2, screen readers, keyboard nav, ARIA, testing, legal compliance)

The English summary is being prepared. Showing the original description.

ziri22/agency-roster6Updated Jul 1, 2026

Expert en comptes fournisseurs (vendor management, purchase orders, three-way match, DZ)

The English summary is being prepared. Showing the original description.

ziri22/agency-roster6Updated Jul 1, 2026

Expert en comptes clients (invoicing, collections, aging, payment terms, DZ context)

The English summary is being prepared. Showing the original description.

ziri22/agency-roster6Updated Jul 1, 2026

See all skills by ziri22

Report a problem with this skill