From dbe562b6c221cd3c75178bfca399dbc4cb01df78 Mon Sep 17 00:00:00 2001 From: Sirius Devops Date: Fri, 15 May 2026 00:31:18 +0000 Subject: [PATCH] =?UTF-8?q?docs:=20comprehensive=20README=20rewrite=20?= =?UTF-8?q?=E2=80=94=20all=20recent=20projects?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Add AgentForge multi-tenant AI agent workspace with dual-tier vLLM - Add Multi-Profile AI Agent Team with automated audit-to-fix pipeline - Add Gateway API / HTTPRoute routing architecture - Add Waitlist API (FastAPI + asyncpg + idempotent signups) - Add CNPG Database Fleet with isolated per-domain clusters - Add Local Business Web Deployment Pipeline - Update AI/ML table: Ollama/Gemma2 → vLLM/Qwen 3.6 27B Abliterated - Update GPU pools: add A100 80GB SPOT deep thinker tier - Update container images: siriussec/* → ghcr.io/sirius0xdev/* - Expand architecture diagram with new namespaces and sec-ops - Expand security section with automated auditing, NetworkPolicy, rate limiting - Remove PineScript note and outdated ollama/siriussec references - Add TOC entry for cost optimization --- README.md | 320 +++++++++++++++++++++++++++++++++++++----------------- 1 file changed, 223 insertions(+), 97 deletions(-) diff --git a/README.md b/README.md index d2a8a31..dc443a0 100644 --- a/README.md +++ b/README.md @@ -1,6 +1,6 @@ # GCloud-Lab DevOps Infrastructure -A cloud-native DevOps laboratory project showcasing modern infrastructure-as-code, GitOps practices, and Kubernetes orchestration on Google Cloud Platform. This project runs a news intelligence system with LLM-powered analysis and a workflow automation platform. +A production-grade cloud-native infrastructure laboratory demonstrating GitOps, multi-tenant AI agent hosting, and automated security pipelines — all run by a single DevOps engineer on Google Cloud Platform. Trusted by builders who ship. ## Table of Contents @@ -13,46 +13,75 @@ A cloud-native DevOps laboratory project showcasing modern infrastructure-as-cod - [Applications](#applications) - [Getting Started](#getting-started) - [Security](#security) +- [Cost Optimization](#cost-optimization) +- [License](#license) --- ## Project Overview -This repository contains infrastructure and application configurations for: +This repository is the single source of truth for a multi-application cloud platform running on GKE. Every deployment, database, and network policy flows through Git via Flux CD. What lives here: -1. **News Intelligence Pipeline**: Automated web scraping, LLM-powered summarization, and Telegram distribution -2. **Workflow Automation**: N8N platform for custom integrations -3. **DevOps Reference Architecture**: Demonstrates GitOps, IaC, and cloud-native best practices +1. **AgentForge** — Private multi-tenant AI agent workspace with dual-tier vLLM inference (L4 dispatcher + A100 deep thinker) and isolated CNPG databases per tenant. +2. **Multi-Profile AI Agent Team** — Six specialist AI profiles (backend-dev, frontend-dev, researcher, outreach, quant, sec-ops) orchestrated through a shared Kanban board with automated audit-to-fix pipelines. +3. **Waitlist API** — FastAPI landing page backend with idempotent signups, async PostgreSQL, and Telegram fire-and-forget notifications. +4. **Autonomous News Quant Pipeline** — 371 global feed scraper with DeepSeek-R1 analysis generating actionable futures trading signals. +5. **N8N Workflow Automation** — Self-hosted workflow engine with dedicated CNPG PostgreSQL. +6. **Local Business Web Deployment Pipeline** — Automated K8s manifest generation for small business websites with cross-namespace HTTPRoute routing. --- ## Architecture ```text -┌─────────────────────────────────────────────────────────────────────────┐ -│ Google Cloud Platform │ -│ ┌───────────────────────────────────────────────────────────────────┐ │ -│ │ GKE Cluster (devops-lab-cluster) │ │ -│ │ │ │ -│ │ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │ │ -│ │ │ Standard │ │ L4 GPU Pool │ │ A100 GPU │ │ │ -│ │ │ Node Pool │ │ (SPOT L4) │ │ (SPOT A100) │ │ │ -│ │ │ e2-std-2 │ │ 1 node (24/7│ │ 0-1 nodes │ │ │ -│ │ └─────────────┘ └─────────────┘ └─────────────┘ │ │ -│ │ │ │ -│ │ ┌─────────────────────────────────────────────────────────────┐ │ │ -│ │ │ Cilium CNI + Hubble │ │ │ -│ │ └─────────────────────────────────────────────────────────────┘ │ │ -│ │ │ │ -│ │ ┌─────────────────────────────────────────────────────────────┐ │ │ -│ │ │ customer1 namespace │ │ │ -│ │ │ - OpenClaw PAaaS (Dual-Tier vLLM: L4 Dispatch / A100 Think)│ │ │ -│ │ │ - News Bot Pipeline (DeepSeek-R1 Quant Analyst) │ │ │ -│ │ │ - PAaaS Landing Page (GHCR Docker Pulls) │ │ │ -│ │ │ - CloudNativePG Isolated Databases │ │ │ -│ │ └─────────────────────────────────────────────────────────────┘ │ │ -│ └───────────────────────────────────────────────────────────────────┘ │ -└─────────────────────────────────────────────────────────────────────────┘ +┌──────────────────────────────────────────────────────────────────────────────┐ +│ Google Cloud Platform │ +│ ┌────────────────────────────────────────────────────────────────────────┐ │ +│ │ GKE Cluster (devops-lab-cluster) │ │ +│ │ │ │ +│ │ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ │ │ +│ │ │ Standard │ │ L4 GPU Pool │ │ A100 GPU Pool│ │ │ +│ │ │ Node Pool │ │ (SPOT L4) │ │ (SPOT A100) │ │ │ +│ │ │ e2-std-2 │ │ 1 node (24/7)│ │ 0-1 nodes │ │ │ +│ │ └──────────────┘ └──────────────┘ └──────────────┘ │ │ +│ │ │ │ +│ │ ┌──────────────────────────────────────────────────────────────────┐ │ │ +│ │ │ Cilium CNI + Hubble + NetworkPolicy │ │ │ +│ │ └──────────────────────────────────────────────────────────────────┘ │ │ +│ │ │ │ +│ │ ┌──────────────────────────────────────────────────────────────────┐ │ │ +│ │ │ Kubernetes Gateway API — external-http-gateway │ │ │ +│ │ │ HTTPRoute PathPrefix → AgentForge / Waitlist / Apps │ │ │ +│ │ └──────────────────────────────────────────────────────────────────┘ │ │ +│ │ │ │ +│ │ ┌───────────────────────────────┐ ┌───────────────────────────────┐ │ │ +│ │ │ customer1 namespace │ │ agent-forge namespace │ │ │ +│ │ │ - AgentForge (PAaaS) │ │ - Tenant-specific OpenClaw │ │ │ +│ │ │ - Dual-tier vLLM │ │ - Isolated CNPG databases │ │ │ +│ │ │ L4 Dispatcher (24/7) │ │ - Qwen 3.6 27B Abliterated │ │ │ +│ │ │ A100 Deep Thinker (KEDA) │ │ - KEDA scale-to-zero │ │ │ +│ │ │ - Waitlist API (FastAPI) │ │ │ │ │ +│ │ │ - News Bot Pipeline │ │ ┌───────────────────────────┐ │ │ │ +│ │ │ - Landing Page │ │ │ sec-ops audit agent │ │ │ +│ │ │ - CNPG PostgreSQL Cluster │ │ │ Automated vuln scanning │ │ │ +│ │ └───────────────────────────────┘ │ │ → backend-dev auto-fix │ │ │ +│ │ │ └───────────────────────────┘ │ │ │ +│ │ ┌───────────────────────────────┐ └───────────────────────────────┘ │ │ +│ │ │ local-business namespaces │ │ │ +│ │ │ - nginx + ConfigMap per biz │ ┌───────────────────────────────┐ │ │ +│ │ │ - Cross-ns HTTPRoute refs │ │ monitoring namespace │ │ │ +│ │ └───────────────────────────────┘ │ - Prometheus + Grafana │ │ │ +│ │ │ - Tailscale-only access │ │ │ +│ │ ┌───────────────────────────────┐ │ - No public ingress │ │ │ +│ │ │ kanban namespace │ └───────────────────────────────┘ │ │ +│ │ │ - Hermes Agent Orchestrator │ │ │ +│ │ │ - 6 Specialist Profiles │ ┌───────────────────────────────┐ │ │ +│ │ │ - Isolated hermes-pgdb │ │ n8n namespace │ │ │ +│ │ └───────────────────────────────┘ │ - Workflow automation │ │ │ +│ │ │ - Dedicated PostgreSQL │ │ │ +│ │ └───────────────────────────────┘ │ │ +│ └────────────────────────────────────────────────────────────────────────┘ │ +└──────────────────────────────────────────────────────────────────────────────┘ ``` --- @@ -92,15 +121,17 @@ This repository contains infrastructure and application configurations for: | Tool | Version | Purpose | |------|---------|---------| | **CloudNative PG** | 0.26.1 | PostgreSQL Kubernetes operator | -| **PostgreSQL** | 15.2 | Relational database (3-node HA cluster) | +| **PostgreSQL** | 15.2 | Relational database (multi-cluster fleet) | ### AI/ML Infrastructure | Tool | Version | Purpose | |------|---------|---------| -| **Ollama** | Latest | Local LLM inference server | -| **Gemma2** | Latest | Open-source LLM for text summarization | -| **NVIDIA L4 GPU** | - | GPU acceleration for LLM workloads | +| **vLLM** | v0.9.1 | High-throughput LLM inference server | +| **Qwen 3.6 27B Abliterated** | Latest | Uncensored reasoning model (A100 deep thinker tier) | +| **Qwen 2.5 Coder 7B Abliterated** | Latest | Fast tool-calling dispatcher (L4 24/7 tier) | +| **NVIDIA L4 GPU** | - | 24/7 GPU for fast triage and dispatch | +| **NVIDIA A100 80GB** | - | SPOT GPU for deep reasoning and multi-file context | ### Development Environment @@ -129,7 +160,7 @@ gcloud-lab/ │ ├── gke.tf # GKE cluster definition │ ├── vpc.tf # VPC and subnet configuration │ ├── nodepool.tf # Standard node pool -│ ├── nodepool-gpu.tf # GPU node pool (SPOT instances) +│ ├── nodepool-gpu.tf # GPU node pools (L4 + A100 SPOT) │ ├── flux.tf # Flux GitOps bootstrap │ ├── helm.tf # Helm chart deployments (Cilium) │ └── variables.tf # Input variables @@ -141,6 +172,7 @@ gcloud-lab/ │ │ ├── gotk-sync.yaml # Git repository sync │ │ └── kustomization.yaml # Flux kustomization │ ├── customer1.yaml # Customer1 Kustomization +│ ├── agent-forge.yaml # AgentForge Kustomization │ ├── infra-controllers.yaml # Infrastructure controllers (CNPG, KEDA, Monitoring, Tailscale) │ └── infra-configs.yaml # Infrastructure configs │ @@ -159,25 +191,48 @@ gcloud-lab/ │ ├── apps/ # Application deployments │ ├── base/ -│ │ └── customer1/ -│ │ ├── namespace.yaml # Namespace definition -│ │ ├── deployment.yaml # N8N deployment -│ │ ├── service.yaml # ClusterIP service -│ │ ├── storage.yaml # PersistentVolumeClaim -│ │ ├── configmap.yaml # N8N configuration -│ │ ├── pg-cluster-customer1.yaml # PostgreSQL cluster -│ │ ├── apigateway.yaml # GCP Gateway -│ │ ├── http-route.yaml # HTTP routing -│ │ ├── healthcheck.yaml # Health check policy -│ │ └── news_bot/ # News bot microservices -│ │ ├── scraper-cronjob.yaml -│ │ ├── analyst-cronjob.yaml -│ │ ├── telebot-cronjob.yaml -│ │ ├── scrapy-configmap.yaml -│ │ └── scrapy-urls-configmap.yaml +│ │ ├── customer1/ +│ │ │ ├── namespace.yaml # Namespace definition +│ │ │ ├── deployment.yaml # N8N + vLLM deployments +│ │ │ ├── service.yaml # ClusterIP services +│ │ │ ├── storage.yaml # PersistentVolumeClaims +│ │ │ ├── configmap.yaml # Application configuration +│ │ │ ├── pg-cluster-customer1.yaml # PostgreSQL cluster +│ │ │ ├── apigateway.yaml # GCP Gateway +│ │ │ ├── http-route.yaml # HTTP routing +│ │ │ ├── healthcheck.yaml # Health check policy +│ │ │ ├── waitlist-api/ # Waitlist API microservice +│ │ │ │ ├── deployment.yaml +│ │ │ │ ├── service.yaml +│ │ │ │ └── configmap.yaml +│ │ │ └── news_bot/ # News bot microservices +│ │ │ ├── scraper-cronjob.yaml +│ │ │ ├── analyst-cronjob.yaml +│ │ │ ├── telebot-cronjob.yaml +│ │ │ ├── scrapy-configmap.yaml +│ │ │ └── scrapy-urls-configmap.yaml +│ │ ├── agent-forge/ +│ │ │ ├── namespace.yaml +│ │ │ ├── vllm-deep-thinker.yaml # A100 deployment with KEDA +│ │ │ ├── openclaw-tenant.yaml # Per-tenant OpenClaw instance +│ │ │ └── pg-cluster-agentforge.yaml +│ │ ├── kanban/ +│ │ │ ├── namespace.yaml +│ │ │ ├── hermes-deployment.yaml # AI agent orchestrator +│ │ │ └── pg-cluster-hermes.yaml +│ │ └── local-business/ +│ │ └── template/ +│ │ ├── namespace.yaml +│ │ ├── nginx-deployment.yaml +│ │ ├── configmap.yaml +│ │ └── http-route.yaml │ └── staging/ -│ └── customer1/ -│ └── kustomization.yaml # Staging overlay +│ ├── customer1/ +│ │ └── kustomization.yaml +│ ├── agent-forge/ +│ │ └── kustomization.yaml +│ └── kanban/ +│ └── kustomization.yaml │ ├── scripts/ │ └── setup # Development setup script @@ -202,8 +257,9 @@ gcloud-lab/ | Pool | Machine Type | Scaling | Purpose | |------|-------------|---------|---------| -| Standard | e2-standard-2 | 1-16 nodes | General workloads | -| GPU (SPOT) | g2-standard-8 + L4 | 0-5 nodes | LLM inference | +| Standard | e2-standard-2 | 1-16 nodes | General workloads, N8N, web servers | +| GPU L4 (SPOT) | g2-standard-8 + L4 | 0-5 nodes | vLLM dispatcher, 24/7 fast inference | +| GPU A100 (SPOT) | a2-highgpu-1g + A100 80GB | 0-1 nodes | Deep thinker tier, multi-file reasoning | ### Networking @@ -211,13 +267,25 @@ gcloud-lab/ - **Primary CIDR**: `10.0.0.0/16` - **Pod CIDR**: `192.168.32.0/20` - **Service CIDR**: `192.168.16.0/24` -- **CNI**: Cilium with advanced datapath -- **Ingress**: GCP L7 Global Load Balancer +- **CNI**: Cilium with advanced datapath and NetworkPolicy enforcement +- **Ingress**: Kubernetes Gateway API via `external-http-gateway` with PathPrefix HTTPRoute routing +- **Internal Services**: Tailscale-only — no public ingress for monitoring, databases, or agent infrastructure + +### CNPG Database Fleet + +Multiple isolated PostgreSQL clusters, each with dedicated databases per application: + +| Cluster | Namespace | Databases | Backup | +|---------|-----------|-----------|--------| +| `customer1-pgdb` | customer1 | `n8n`, `news_app`, `waitlist` | GCS, 7-day retention | +| `hermes-pgdb` | kanban | `hermes`, `memory_store` | GCS, 7-day retention | +| `openclaw-pgdb` | agent-forge | Per-tenant isolated DBs | GCS, 7-day retention | +| `siriusdevops-pgdb` | customer1 | `waitlist_prod` | GCS, 30-day retention | ### GitOps Flow ``` -GitHub Repository +GitHub Repository (ghcr.io/sirius0xdev) │ ▼ Flux Source Controller (watches git, 1min interval) @@ -227,31 +295,76 @@ GitHub Repository │ ├── infrastructure/controllers → CNPG, KEDA, Monitoring, Tailscale ├── infrastructure/configs → Cluster configs - └── apps/staging/customer1 → Applications + ├── apps/staging/customer1 → PAaaS, N8N, News Bot, Waitlist API + ├── apps/staging/agent-forge → Multi-tenant AI agent hosting + ├── apps/staging/kanban → AI Agent Team orchestrator + └── apps/staging/local-business → Business websites ``` --- ## Applications -### 1. Private Assistant as a Service (PAaaS) -A premium, uncensored, privacy-first AI assistant platform with dual-tier cognitive architecture: -- **Tier 1 (Dispatcher):** L4 GPU Spot instance running 24/7. Hosts `Qwen2.5-Coder-7B-Instruct-heretic` via vLLM `v0.9.1` for lightning-fast, cheap triage and tool calling (using the `pythonic` tool parser). -- **Tier 2 (Deep Thinker):** A100 80GB Spot instance scaling from 0-1 via KEDA. Hosts `Qwen3.5-27B-heretic` with `--enable-chunked-prefill` and `--kv-cache-dtype=fp8` for massive multi-file context and reasoning without OOMing or stalling concurrent users. -- **Frontend:** Isolated `openclaw` deployments per tenant, connected to Telegram/Discord via outbound polling (no public ingress required). -- **Landing Page:** Dockerized marketing site built via CI/CD from `openclaw-projects` and deployed to the `staging` kustomization overlay. +### 1. AgentForge — Private AI Agent Workspace + +A premium, uncensored, privacy-first AI agent hosting platform with dual-tier cognitive architecture: + +- **Tier 1 (Dispatcher):** L4 GPU SPOT instance running 24/7. Hosts `Qwen2.5-Coder-7B-Instruct-heretic` via vLLM `v0.9.1` for lightning-fast, cheap triage and tool calling. +- **Tier 2 (Deep Thinker):** A100 80GB SPOT instance scaling from 0-1 via KEDA. Hosts `Qwen3.6-27B-heretic` with chunked prefill and FP8 KV cache for massive multi-file context and reasoning without OOMing. +- **Multi-Tenant Isolation:** Each tenant gets an isolated OpenClaw deployment with its own CNPG PostgreSQL database. No cross-tenant data leakage. +- **Landing Page:** Dockerized marketing site at siriusdevops.com, built via CI/CD from GitHub Actions and deployed to the staging kustomization overlay. +- **Container Registry:** All images pushed to `ghcr.io/sirius0xdev`. + +### 2. Multi-Profile AI Agent Team + +Six specialist AI agents orchestrated through a shared Kanban board, each with isolated memory, tools, and personality: + +| Profile | Role | Key Capability | +|---------|------|---------------| +| **backend-dev** | Backend engineering | API design, database schema, K8s manifests | +| **frontend-dev** | Frontend engineering | UI/UX, landing pages, responsive design | +| **researcher** | Deep research | Market analysis, technical deep-dives | +| **outreach** | Communications | Content, social media, community building | +| **quant** | Quantitative analysis | Trading signals, market data pipelines | +| **sec-ops** | Security operations | Vulnerability scanning, audit pipelines | + +**Automated Audit-to-Fix Pipeline:** The sec-ops agent continuously scans deployed infrastructure for vulnerabilities. When findings are confirmed, the backend-dev agent is automatically dispatched to remediate — from detection to patch in a single GitOps cycle. + +### 3. Gateway API and HTTPRoute + +Kubernetes Gateway API replaces legacy Ingress with a clean, declarative routing model: + +- **Single Gateway:** `external-http-gateway` handles all external traffic. +- **PathPrefix Routing:** `/agentforge/*` → AgentForge landing, `/waitlist/*` → Waitlist API, `/business/*` → local business sites. +- **No Public Ingress for Internals:** Monitoring (Grafana/Prometheus), databases, and agent infrastructure are accessible only via Tailscale VPN. +- **Cross-Namespace References:** HTTPRoute resources in one namespace can reference Services in another, keeping routing centralized. + +### 4. Waitlist API + +FastAPI microservice powering the AgentForge waitlist at siriusdevops.com: + +- **Database:** asyncpg connection pool to dedicated CNPG PostgreSQL. +- **Idempotent Signups:** `INSERT ... ON CONFLICT DO NOTHING` — duplicate emails are silently ignored, not rejected. +- **Notifications:** Fire-and-forget Telegram webhook on each new signup. No blocking I/O in the request path. +- **Security:** Rate limiting per IP, input sanitization, and CORS whitelist. + +### 5. Autonomous News Quant Pipeline (`news_bot`) -### 2. Autonomous News Quant Pipeline (`news_bot`) An institutional-grade pipeline scraping 371 global feeds to generate actionable futures trading signals: + - **Scraper:** CronJob at `:50` pulling multi-lingual global financial data. -- **Map/Reduce Analyst:** Utilizes DeepSeek-R1 (with a strict 10-step `` protocol) and local open-weights to extract "Market-Moving DNA". Translates events into explicit futures targets (/ES, /CL, /NQ) with R:R, Take Profit, and Stop Loss levels anchored in provided volume/price data. -- **Privacy:** All proprietary technical data stays strictly within the VPC, executing against local models rather than public APIs like OpenAI to protect the trading edge and avoid throttling during market panics. +- **Map/Reduce Analyst:** DeepSeek-R1 with a strict 10-step think protocol extracts "Market-Moving DNA" and translates events into explicit futures targets (/ES, /CL, /NQ) with risk:reward, take profit, and stop loss levels. +- **Privacy:** All proprietary technical data stays strictly within the VPC, executing against local models to protect the trading edge. -### 3. N8N Workflow Automation -- **Database**: PostgreSQL (dedicated `n8n` database) -- Custom integrations and webhook catchers. +### 6. Local Business Web Deployment Pipeline -*(Note: PineScript trading strategies have been migrated out of this IaC repository and live in `openclaw-projects/trading-bots`.)* +Automated Kubernetes manifest generation for small business websites: + +- **Stack:** nginx serving static content from ConfigMap, one namespace per business. +- **Routing:** HTTPRoute with cross-namespace Service references under `/business/` paths. +- **Zero Cold Start:** Static sites have no database dependency — just nginx + ConfigMap, deployed via GitOps. + +--- ## Getting Started @@ -342,16 +455,48 @@ or create a `monitoring-grafana-admin` Secret instead. ### Network Security -- Cilium network policies for pod-to-pod isolation -- TLS termination at load balancer -- Private cluster networking with NAT +- Cilium NetworkPolicy for pod-to-pod and namespace-to-namespace isolation +- Kubernetes Gateway API with TLS termination at the load balancer +- Internal services (monitoring, databases, agent infrastructure) accessible only via Tailscale VPN — zero public ingress +- Rate limiting on public-facing APIs (Waitlist, landing page) ### Database Security -- Managed roles with secret-based passwords -- Separate users per application (`customer1`, `news_app`) +- Managed roles with secret-based passwords per application +- Separate PostgreSQL clusters per domain (hermes-pgdb, openclaw-pgdb, siriusdevops-pgdb) +- GCS backups with configurable retention policies - HA cluster with automatic failover +### Automated Security Auditing + +- **sec-ops Agent:** Continuously scans deployed infrastructure for CVEs, misconfigurations, and policy violations +- **Auto-Remediation:** Confirmed findings automatically dispatch the backend-dev agent to patch and commit +- **Audit Trail:** Every finding, fix, and deployment is tracked in Git history — full provenance from detection to resolution + +--- + +## Cost Optimization + +- **SPOT GPU Instances**: 60-90% savings on L4 and A100 workloads +- **KEDA Scale-to-Zero**: A100 deep thinker pool scales to 0 when no requests are queued +- **Resource Limits**: CPU and memory caps on every container prevent runaway costs +- **Scheduled Workloads**: CronJobs only run when needed — no idle inference pods +- **Tailscale for Internal Access**: No need for expensive internal load balancers or Cloud NAT for monitoring + +--- + +## Container Images + +``` +ghcr.io/sirius0xdev/agentforge-landing:latest +ghcr.io/sirius0xdev/waitlist-api:latest +ghcr.io/sirius0xdev/newsscraper:latest +ghcr.io/sirius0xdev/summarizer:latest +ghcr.io/sirius0xdev/news-messenger:latest +docker.n8n.io/n8nio/n8n:2.1.4 +ghcr.io/cloudnative-pg/postgresql:15.2 +``` + --- ## Tool Reference @@ -369,30 +514,11 @@ flux = "~> 1.7" # GitOps bootstrap ```yaml cilium: 1.18.5 # CNI and service mesh cloudnative-pg: 0.26.1 # PostgreSQL operator +vllm: 0.9.1 # High-throughput LLM serving ``` -### Container Images - -``` -docker.n8n.io/n8nio/n8n:2.1.4 -ghcr.io/cloudnative-pg/postgresql:15.2 -ollama/ollama:latest -siriussec/newsscraper:latest -siriussec/summarizer:latest -siriussec/news-messenger:latest -``` - ---- - -## Cost Optimization - -- **SPOT GPU Instances**: 60-90% savings on LLM workloads -- **Autoscaling**: GPU nodes scale to 0 when idle -- **Resource Limits**: Prevents runaway costs -- **Scheduled Workloads**: CronJobs only run when needed - --- ## License -Private repository - All rights reserved. +Private repository — All rights reserved.