scale-ingestion-service
Adjust RL ingestion service worker count based on queue backlog to maintain optimal throughput.
Files12
View on GitHubSKILL.mdskills/scale-ingestion-service/SKILL.md
GitHubscale-ingestion-service Specification
ID: scale-ingestion-service
Version: 1.0.0
Status: Active
Owner: CIC Pipeline Team
Purpose
Monitor ingestion queue depth and automatically scale workers:
- Queue > 500: Scale to 8 workers + send alert
- Queue 100–500: Scale to 4 workers
- Queue < 100: Scale to 2 workers
Trigger
scale-ingestion-service [target-queue-depth]
Input Schema
interface SkillInput {
targetQueueDepth?: number; // Optional override (100-500; default: 100)
}
Output Schema
interface SkillOutput {
status: "success" | "error";
currentQueueDepth: number;
recommendedWorkers: number;
previousWorkers: number;
deploymentTime: string;
timestamp: string;
}
Error Handling
See Skill Operator Guide — Error Handling for standard error codes.
Additional errors:
| Code | Message | Handler |
|---|---|---|
QUEUE_UNAVAILABLE |
Cannot access queue service | Check service connectivity |
DEPLOYMENT_FAILED |
Worker scaling failed | Retry or manual rollout |
Full Reference
For Setup, Requirements, Testing, Configuration, and Troubleshooting:
→ See Skill Operator Guide