Kalmitech Startups / Production Showcase
portfolio.kalmitech.app Operator Sign In
Kalmitech Logo
Kalmitech Technology Studio
Production SaaS Platforms DevOps & Infrastructure Featured

KalmiDeploy

Self-hosted platform-as-a-service for managing and deploying containerized applications on Linux VPS environments.

Release Year
2025
Entity / Scope
Kalmitech Technologies
Engineering Role
Lead Systems Architect
Lifecycle State
Production
Launch Application Repository Documentation
KalmiDeploy Hero Screenshot
Overview

System Architecture & Purpose

KalmiDeploy is a lightweight orchestration dashboard designed to build, deploy, and monitor multi-container web stacks directly on independent Linux VPS instances without commercial cloud vendor lock-in. The system interfaces directly with the host Docker socket, provides real-time CPU, memory, network, and disk I/O metrics, automates zero-downtime rolling builds from Git repositories, streams deployment logs via Server-Sent Events, dynamically updates Nginx reverse-proxy upstreams, and runs automated repository health audits.
The Challenge

Problem Statement

Deploying multiple web services across Linux virtual private servers historically required manual SSH interventions, manual Nginx configuration edits, risky database migration steps, and unmonitored container drift. Commercial cloud PaaS options introduced high monthly per-container costs, while Kubernetes introduced excessive overhead for single-node and multi-service setups.
Engineering Resolution

Solution Design

KalmiDeploy provides an autonomous management layer directly on the host. It inspects Docker Compose definitions, orchestrates decoupled builds using Celery async queues, verifies container network attachments before traffic routing, and performs automated client-side encrypted database backups to off-site storage.
Capabilities

Key Features & Architecture

1

Fleet Telemetry

Real-time CPU, memory, network I/O, block I/O, and PID monitoring across all active container tiers.

2

Zero-Downtime Rollouts

Automated git checkout, multi-stage image builds, and container replacement with live streaming logs.

3

Dynamic Reverse-Proxy Sync

Automatic generation and reloading of Nginx upstream configurations on service renaming.

4

Pre-Flight Audits

Automated repository audits inspecting twelve-factor settings, Dockerfiles, Compose networks, and environment variables.

5

Real-Time Log Streaming

Server-Sent Events streaming build and container logs to operator consoles in real time.

6

Encrypted Backups

Client-side encrypted nightly PostgreSQL backups automatically synchronized to off-site cloud storage.

Infrastructure

Architecture & System Topology

Client Requests -> Master Nginx Reverse Proxy (SSL Termination, HTTP/2, Gzip) -> KalmiDeploy Control Plane (Django / Gunicorn) * Docker Socket Integration (/var/run/docker.sock) for container fleet management * Celery Worker Queues & Redis Pub/Sub for background builds and log streaming * Dynamic Nginx Upstream Synchronizer for atomic configuration reloads * Backup Synchronizer for encrypted off-site database archives
COMPOSE SERVICE DISCOVERY

Compose Service Discovery

Resolving container drift and orphan service targeting

Traditional server scripts often rely on static container names. When an application updates its Compose service name or image tag, scripts fail or target stopped containers. KalmiDeploy inspects compose file specifications and container labels to identify running service instances, verify network attachment to the web proxy network, and test port readiness before routing live web traffic.
DISTRIBUTED CONCURRENCY CONTROLS

Distributed Concurrency Controls

Eliminating race conditions during rapid deployments

When multiple webhooks trigger builds simultaneously, concurrent compilation can exhaust host memory. KalmiDeploy enforces Redis-backed distributed mutex locks on all deployment pipelines. Subsequent triggers are queued sequentially in Celery worker queues, guaranteeing predictable execution and resource stability.
Screenshots & Interface Walkthrough

Interface Documentation

Visual overview of production interfaces, management consoles, and operational workflows. Click any screenshot for high-resolution inspection.

Cluster Observability & Container Fleet: Real-time telemetry monitoring CPU load, memory utilization, NVMe storage, and unified health statuses for 14 active microservices.
Cover Thumbnail
Figure 01

Cluster Observability & Container Fleet: Real-time telemetry monitoring CPU load, memory utilization, NVMe storage, and unified health statuses for 14 active microservices.

Production Application Catalog: Microservices fleet status dashboard showing reverse-proxy port allocations, container state indicators, and direct web ingress links.
Figure 02

Production Application Catalog: Microservices fleet status dashboard showing reverse-proxy port allocations, container state indicators, and direct web ingress links.

Service Topology & Runtime Controls: Microservice inspection console with dynamic container health checks, autofix triggers, live restart hooks, and environment variables.
Figure 03

Service Topology & Runtime Controls: Microservice inspection console with dynamic container health checks, autofix triggers, live restart hooks, and environment variables.

Persistent Storage & Cloud Backup Manager: Automated PostgreSQL and volume snapshot scheduler with encrypted remote Rclone bucket replication and point-in-time recovery.
Figure 04

Persistent Storage & Cloud Backup Manager: Automated PostgreSQL and volume snapshot scheduler with encrypted remote Rclone bucket replication and point-in-time recovery.

GitHub CI/CD Deployment Integration: Automated webhook repository tracker triggering headless build pipelines, container rebuilding, and zero-downtime rolling deploys.
Figure 05

GitHub CI/CD Deployment Integration: Automated webhook repository tracker triggering headless build pipelines, container rebuilding, and zero-downtime rolling deploys.

Outcomes & Results

Delivery Impact

Powers internal production services across Kalmitech VPS infrastructure, eliminating manual deployment errors, reducing release cycle times to under 30 seconds, and ensuring high uptime with zero licensing fees.