Deployment

This page covers running Blink Store in production: standalone, Docker, resource limits, and logging.

---

Standalone

Install the latest binary (see Installation) and run:

./blink-store serve --tcp 0.0.0.0:8765 --memory-limit 10485760

Flag	Default	Description
--tcp <addr>:<port>	(none)	TCP listen address. Use 0.0.0.0 to accept external connections. At least one of --tcp or --unix is required.
--unix <path>	(none)	Unix domain socket path (Linux/macOS only).
--memory-limit <bytes>	10485760 (10 MiB)	Maximum total size of stored keys + values. LRU eviction kicks in when exceeded.
--log-dir <path>	(none)	Directory for rolling log files.
--retention-minutes <n>	(none)	Auto-delete log files older than n minutes.

---

Docker

Quick start

docker run -d --name blink-store \
  -p 8765:8765 \
  ghcr.io/ashokdudhade/blink-store:latest

Override defaults with environment variables:

docker run -d --name blink-store \
  -p 9000:9000 \
  -e BLINK_PORT=9000 \
  -e BLINK_MEMORY_LIMIT=104857600 \
  -e BLINK_RETENTION_MINUTES=1440 \
  ghcr.io/ashokdudhade/blink-store:latest

Or build a local image (downloads the pre-built binary from GitHub Releases — no Rust toolchain needed):

curl -sSLf -o Dockerfile \
  https://raw.githubusercontent.com/ashokdudhade/blink-store/main/Dockerfile
docker build --build-arg BLINK_VERSION=latest -t blink-store .
docker run -d -p 8765:8765 blink-store

Docker Compose

services:
  blink-store:
    image: ghcr.io/ashokdudhade/blink-store:latest
    ports:
      - "8765:8765"
    environment:
      - BLINK_MEMORY_LIMIT=10485760
    deploy:
      resources:
        limits:
          memory: 32M
          cpus: "0.5"
        reservations:
          memory: 8M
    restart: unless-stopped

docker compose up -d

---

Resource limits

Memory

Layer	Flag / setting	Purpose
Application	--memory-limit <bytes>	Caps total stored data. Triggers LRU eviction.
Container	deploy.resources.limits.memory	Hard cap on the container. Prevents OOM-kill from affecting the host.

Set the container limit higher than --memory-limit to account for process overhead (stack, buffers, etc.). A ratio of 3:1 is a safe starting point — e.g. --memory-limit 10485760 (10 MiB) with a container limit of 32M.

CPU

deploy:
  resources:
    limits:
      cpus: "0.5"

Or use cpuset for core pinning in performance-sensitive scenarios.

---

Logging

Enable structured logging to files with automatic rotation:

./blink-store serve --tcp 0.0.0.0:8765 \
  --memory-limit 10485760 \
  --log-dir /var/log/blink-store \
  --retention-minutes 1440

Flag	Description
--log-dir	Directory for rolling log files. Created if it doesn't exist.
--retention-minutes	Prune log files older than this. 1440 = 24 hours.

Logs use structured tracing format and include timestamps, log levels, key names, and byte sizes for every operation.

---

Health check

Blink Store responds to USAGE with the current memory usage:

echo "USAGE" | nc 127.0.0.1 8765
# → USAGE 1234

Use this as a TCP health check in Docker, Kubernetes, or your load balancer. A response confirms the server is alive and accepting connections.

---

Production checklist

• Set --memory-limit based on available RAM
• Set container memory limits (Docker / K8s)
• Enable --log-dir with --retention-minutes
• Bind to 0.0.0.0 only if external access is needed; use 127.0.0.1 for local-only
• Pin the binary version (e.g. v0.1.0) in production deployments instead of latest
• Add a health check on the USAGE command