SERVICE COMPARISON

Service Comparison and Decision Matrix

Practical guide to choosing which optional services to enable in your nself deployment. Required services (PostgreSQL, Hasura, Auth, Nginx) are always running -- this page focuses on the decisions you actually need to make.

---

Decision Matrix

Use this table to determine which optional services your project needs.

Service	What It Does	Enable When	Skip When
Redis	In-memory cache, session store, pub/sub messaging	You need session management, caching, background jobs (BullMQ), or horizontal scaling	Single-instance app using only JWT auth with no caching layer
MinIO	S3-compatible object storage for files and media	You handle file uploads, user avatars, documents, or need artifact storage for MLflow	Text-only application or you use an external S3 provider (AWS, GCP, Backblaze)
nself-admin	Web UI for managing your entire nself deployment	You want a visual dashboard for service health, config management, and log viewing	You prefer CLI-only workflows or are running in headless CI/CD environments
Functions	Serverless JavaScript/TypeScript runtime	You need lightweight endpoints, webhooks, or event-driven logic without a full custom service	You already have custom services (CS_N) handling all your backend logic
Mail	Email sending and testing (MailPit in dev, SMTP in prod)	Your app sends transactional email: signups, password resets, notifications	No email functionality needed, or you handle email entirely in a custom service
Search	Full-text search with typo tolerance (MeiliSearch)	You need instant search, faceted filtering, or autocomplete across your data	Simple queries handled by PostgreSQL full-text search are sufficient
MLflow	ML experiment tracking, model registry, artifact storage	You train models, run experiments, or need a model registry	No machine learning workloads in your project
Monitoring	Full observability: metrics, logs, traces, alerts (10 services)	Production deployment, staging validation, or debugging performance issues	Local development only, or you use an external monitoring provider (Datadog, New Relic)

---

Resource Impact

Approximate resource consumption per service under typical workloads. Actual usage varies with traffic and data volume.

Service	RAM (Idle)	RAM (Active)	CPU (Avg)	Disk	Has Public Route
PostgreSQL	128 MB	256-512 MB	0.25 cores	1-50 GB	No (internal)
Hasura	128 MB	256 MB	0.25 cores	Minimal	Yes
Auth	64 MB	128 MB	0.1 cores	Minimal	Yes
Nginx	32 MB	64 MB	0.1 cores	Minimal	Yes (gateway)
Redis	64 MB	256 MB	0.25 cores	1 GB	No (internal)
MinIO	256 MB	512 MB	0.5 cores	Varies	Yes
nself-admin	128 MB	256 MB	0.25 cores	100 MB	Yes
Functions	128 MB	256 MB	0.25 cores	500 MB	Yes
Mail (MailPit)	64 MB	128 MB	0.1 cores	100 MB	Yes
Search (MeiliSearch)	256 MB	512 MB	0.5 cores	1-5 GB	Yes
MLflow	256 MB	512 MB	0.5 cores	1-10 GB	Yes
Monitoring Bundle	1.5 GB	3.5 GB	2.5 cores	10-26 GB	Yes (3 routes)

Monitoring Bundle Breakdown

The monitoring bundle enables 10 services as a group. Individual resource usage:

Monitoring Service	RAM	CPU	Disk	Purpose
Prometheus	512 MB	0.5 cores	5-10 GB	Metrics collection and storage
Grafana	256 MB	0.25 cores	500 MB	Visualization dashboards
Loki	256 MB	0.25 cores	2-5 GB	Log aggregation
Promtail	64 MB	0.1 cores	Minimal	Log shipping to Loki
Tempo	128 MB	0.25 cores	1-5 GB	Distributed tracing
Alertmanager	64 MB	0.1 cores	Minimal	Alert routing
cAdvisor	128 MB	0.5 cores	Minimal	Container metrics
Node Exporter	32 MB	0.1 cores	Minimal	System metrics
Postgres Exporter	32 MB	0.1 cores	Minimal	Database metrics
Redis Exporter	16 MB	0.05 cores	Minimal	Redis metrics

---

Recommended Configurations

Simple Blog or Website

Required services only. PostgreSQL stores your content, Hasura provides the API, Auth handles logins, Nginx routes traffic.

• Containers: 4
• RAM: ~400 MB
• Recommended minimum: 1 GB RAM, 1 vCPU

# .env - No optional services needed
PROJECT_NAME=myblog
ENV=dev
BASE_DOMAIN=localhost
POSTGRES_PASSWORD=secure-password
HASURA_GRAPHQL_ADMIN_SECRET=admin-secret

SaaS Application

Add Redis for session management and caching, MinIO for file storage, and Mail for transactional email.

• Containers: 7
• RAM: ~1.2 GB
• Recommended minimum: 2 GB RAM, 2 vCPU

# .env - SaaS essentials
PROJECT_NAME=myapp
ENV=dev
BASE_DOMAIN=localhost
POSTGRES_PASSWORD=secure-password
HASURA_GRAPHQL_ADMIN_SECRET=admin-secret

REDIS_ENABLED=true
MINIO_ENABLED=true
MAILPIT_ENABLED=true

E-Commerce Platform

Everything in SaaS, plus Search for product catalog and browse/filter functionality.

• Containers: 8
• RAM: ~1.7 GB
• Recommended minimum: 4 GB RAM, 2 vCPU

# .env - E-commerce stack
PROJECT_NAME=store
ENV=dev
BASE_DOMAIN=localhost
POSTGRES_PASSWORD=secure-password
HASURA_GRAPHQL_ADMIN_SECRET=admin-secret

REDIS_ENABLED=true
MINIO_ENABLED=true
MAILPIT_ENABLED=true
MEILISEARCH_ENABLED=true

Enterprise / Multi-Tenant

All 7 optional services enabled. Full platform with admin UI, storage, cache, search, email, functions, and ML tracking.

• Containers: 11
• RAM: ~2.5 GB
• Recommended minimum: 4 GB RAM, 4 vCPU

# .env - Enterprise configuration
PROJECT_NAME=platform
ENV=dev
BASE_DOMAIN=localhost
POSTGRES_PASSWORD=secure-password
HASURA_GRAPHQL_ADMIN_SECRET=admin-secret

REDIS_ENABLED=true
MINIO_ENABLED=true
NSELF_ADMIN_ENABLED=true
FUNCTIONS_ENABLED=true
MAILPIT_ENABLED=true
MEILISEARCH_ENABLED=true
MLFLOW_ENABLED=true

Full Demo (Everything)

All optional services, full monitoring bundle, plus 4 custom services. This is what nself init --demo generates.

• Containers: 25
• RAM: ~6 GB
• Recommended minimum: 8 GB RAM, 4 vCPU

# .env - Full demo
PROJECT_NAME=demo-app
ENV=dev
BASE_DOMAIN=localhost
POSTGRES_PASSWORD=secure-password
HASURA_GRAPHQL_ADMIN_SECRET=admin-secret

# All optional services
REDIS_ENABLED=true
MINIO_ENABLED=true
NSELF_ADMIN_ENABLED=true
FUNCTIONS_ENABLED=true
MAILPIT_ENABLED=true
MEILISEARCH_ENABLED=true
MLFLOW_ENABLED=true

# Monitoring bundle (10 services)
MONITORING_ENABLED=true

# Custom services (4)
CS_1=express-api:express-js:8001
CS_2=bullmq-worker:bullmq-js:8002
CS_3=grpc-api:grpc:8003
CS_4=ml-api:fastapi:8004

---

Port Reference

Default ports for all services. Internal ports are used inside the Docker network. External ports are what you access from the host.

Required Services

Service	Internal Port	External Route	Protocol
PostgreSQL	5432	None (use POSTGRES_PORT for host access)	TCP
Hasura	8080	api.<domain>	HTTPS
Auth	4000	auth.<domain>	HTTPS
Nginx	80, 443	<domain> (gateway)	HTTP/HTTPS

Optional Services

Service	Internal Port	External Route	Env Variable
Redis	6379	None (internal only)	REDIS_ENABLED=true
MinIO API	9000	minio.<domain>	MINIO_ENABLED=true
MinIO Console	9001	minio.<domain>/console	(same)
nself-admin	3021	admin.<domain>	NSELF_ADMIN_ENABLED=true
Functions	3008	functions.<domain>	FUNCTIONS_ENABLED=true
MailPit SMTP	1025	None (internal SMTP)	MAILPIT_ENABLED=true
MailPit UI	8025	mail.<domain>	(same)
MeiliSearch	7700	search.<domain>	MEILISEARCH_ENABLED=true
MLflow	5005	mlflow.<domain>	MLFLOW_ENABLED=true

Monitoring Bundle

Service	Internal Port	External Route	Can Disable
Prometheus	9090	prometheus.<domain>	Not recommended
Grafana	3000	grafana.<domain>	Not recommended
Loki	3100	None (internal)	Not recommended
Promtail	9080	None (internal)	Required for Loki
Tempo	3200, 14268	None (internal)	TEMPO_ENABLED=false
Alertmanager	9093	alertmanager.<domain>	ALERTMANAGER_ENABLED=false
cAdvisor	8081	None (internal)	CADVISOR_ENABLED=false
Node Exporter	9100	None (internal)	NODE_EXPORTER_ENABLED=false
Postgres Exporter	9187	None (internal)	POSTGRES_EXPORTER_ENABLED=false
Redis Exporter	9121	None (internal)	REDIS_EXPORTER_ENABLED=false

---

Configuration Summary by Container Count

Profile	Required	Optional	Monitoring	Custom	Total	Min RAM
Simple blog	4	0	0	0	4	1 GB
SaaS app	4	3	0	0	7	2 GB
E-commerce	4	4	0	0	8	4 GB
Enterprise	4	7	0	0	11	4 GB
Prod + monitoring	4	3	10	0	17	4 GB
Full demo	4	7	10	4	25	8 GB

Start with the fewest services you need. You can enable additional services at any time by adding the corresponding *_ENABLED=true variable to your .env file and running nself build && nself restart.

---

Related Documentation

• Services Overview -- All service categories
• Optional Services -- Detailed optional service configuration
• Monitoring Bundle -- Full monitoring stack reference
• Custom Services -- Building your own services with CS_N
• Demo Setup -- Full 25-container demo configuration
• Start Command Options -- Tuning startup behavior

---

Back to Services Index | Back to Documentation Home