Skip to main content
Open sourceDocker ComposeLocal AI

Observe every signal. Investigate with AI.

A composable observability and AIOps platform for Docker hosts, microservices, and local LLM workloads—with metrics, logs, traces, MCP endpoints, runbook retrieval, and Generative UI in one stack.

$ docker compose up -d
$ docker compose ps

prometheus   running
grafana      running
loki         running
tempo        running
Unified telemetryPrometheus, Grafana, Loki, Tempo, and OpenTelemetry.
AIOps interfacesGrafana and Prometheus MCP endpoints for operational tooling.
Local-first copilotOllama, Chainlit, OpenUI, and pgvector remain on your network.

Architecture overview​

Dockprom-AI operates as a coordinated dual-stack architecture communicating across a high-isolation bridge network (monitor-net):

  1. Core observability and AIOps stack: telemetry ingest, time-series storage, tracing, log aggregation, alerting pipelines, and MCP endpoints.
  2. Generative UI and copilot stack (compose-genui.yaml): conversational workflows, vector search over runbooks, and interactive remediation interfaces.

Core observability stack​

The core pipeline captures metrics, distributed traces, and container logs with deep Linux kernel and host hardware instrumentation:

DOCKER HOST PLATFORMLinuxOllama Gateway (Envoy)Port :11435 • v1.40.2OTLP Traces & RPS MetricsHost Native Ollama0.0.0.0:11434 • Host GPULlama 3.2 • nomic-embed-textAI Coding AgentsClaude Desktop • CursorCline • Roo Code • ChainlitSSE Transport (v2.1)Grafana MCP ServerPort :8000/sse • v2.1.1Prometheus MCP ServerPort :8080/sse • v0.18.0Caddy v2 Proxy:3000, :9090, :9093Bcrypt Basic Authv2.11.7-alpineGrafana 13.3.0Port :3000 • DashboardsExplore • LogQL & TraceQLPrometheus v3.15.1Port :9090 • PromQL TSDBHead Compaction • 200h RetAlertmanager v0.34.1Port :9093 • DeduplicationWebhooks & RoutingDocker Workloads & HostsLogporter v1.5.0:9333 • Container CPU/Mem/LogsNode Exporter v1.12.1:9100 • Linux Kernel/FS/IOOTel Collector Contribv0.164.0OTLP Ingest :4317 / :4318Pipelines: Traces • Logs • MetricsProm Exp :8889Tempo v3.1.1:3200 • Distributed TracingWAL & Parquet storageLoki v3.7.9:3100 • TSDB v13 SchemaStructured metadataKeep AIOps v0.54.3UI :8082 • API :8080Alert Enrichment & CorrelationMattermost Team 12.0.1Port :8065 • Incident ChannelsAutomated Webhooks & ChatOpsPostgreSQL 17 + pgvector (Unified Database)Container: mattermost-db • Port :5432 • pgvector/pgvector:pg17Keep AIOps Storedb: keep • alert eventsMattermost Storagedb: mattermost • users/chatRAG Vector Embeddingsdb: rag_storage • vector(768) HNSWpostgres-exporter v0.20.1:9187 • pg_stat_statementsBuffer hits • locks • latency

Generative UI and copilot pipeline​

The GenUI pipeline connects user prompts to real-time telemetry, semantic runbook embeddings, and streaming React components:

GENUI & AIOPS COPILOT STACKcompose-genuiSRE Operator"Diagnose CPU load &render remediation tool"Chainlit AIOps CopilotPort :8050 • chainlit-genuiStep 1: pgvector RAG LookupEmbedding cosine similarityStep 2: MCP Tool InvocationQuery PromQL & Loki logsStep 3: Streaming ReasoningOpenUI Lang component streamInteractive Preview iframe Embedpgvector RAG Store (PostgreSQL 17)Database: rag_storage • Table: operational_runbooksHNSW Cosine Vector Index (<=>)vector(768) • Matches nomic-embed-textObservability MCP ServersHTTP / Server-Sent Events (SSE)Grafana MCP:8000/ssePrometheus MCP:8080/sseEnvoy Gateway & Host Ollamahttp://ollama-gateway:11435 -> 0.0.0.0:1143432k Context Window (OLLAMA_CONTEXT_LENGTH=32768)Llama 3.2 • nomic-embed-text • Streamed TokensOpen-Intelligent-UI CanvasPort :3001 • Container: open-intelligent-ui • Next.js + @openuidev/reactContainer CPU Usage92.4%Threshold ExceededSuggested SRE ActionRestart high-consumer container:⚡ Restart ContainerLive PromQL Telemetryrate(container_cpu[1m])

Quickstart​

The GenUI stack assumes the core services and external monitor-net network already exist.

export PGVECTOR_URL='postgresql://<user>:<password>@mattermost-db:5432/rag_storage'
docker compose -f compose-genui.yaml config --quiet
docker compose -f compose-genui.yaml up -d --build

Open Chainlit at http://localhost:8050 and Open-Intelligent-UI at http://localhost:3001.

Continue with the host prerequisites, environment reference, and GenUI architecture.