惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

G
Google Developers Blog
F
Fortinet All Blogs
Microsoft Azure Blog
Microsoft Azure Blog
腾讯CDC
Vercel News
Vercel News
Recent Announcements
Recent Announcements
博客园 - Franky
小众软件
小众软件
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
The Cloudflare Blog
宝玉的分享
宝玉的分享
I
InfoQ
博客园 - 聂微东
Jina AI
Jina AI
J
Java Code Geeks
V
V2EX
U
Unit 42
Stack Overflow Blog
Stack Overflow Blog
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
阮一峰的网络日志
阮一峰的网络日志
L
LangChain Blog
T
The Blog of Author Tim Ferriss
量子位

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
Why Kubernetes Needs an IDP, Not Just GitOps
Dima S · 2026-05-03 · via DEV Community

We had ArgoCD running perfectly. Every deployment was reconciled from Git. Drift detection worked. Rollbacks were one-click. Our GitOps setup was clean.

Developers still couldn't provision a staging environment without pinging the platform team.

That gap — between "GitOps in place" and "developers can actually self-serve" — is where most platform engineering teams get stuck. GitOps solves a real problem. It just doesn't solve all the problems people think it solves.

What GitOps Actually Solves

Let's be precise, because GitOps deserves credit where it's due.

ArgoCD (17k+ GitHub stars) and Flux are excellent tools for one specific job: keeping your cluster state in sync with what's declared in Git. They watch a repository, detect drift between desired and actual state, and reconcile continuously.

This is genuinely valuable:

  • Deployment consistency — no more "works on my machine, breaks in prod" config drift
  • Rollback — revert a deployment by reverting a commit
  • Audit trail for deployments — Git history shows who changed what
  • Multi-cluster sync — apply the same manifests across clusters reliably

If you're not using GitOps, you should be. It's the right tool for deployment reconciliation.

The problem is what it doesn't do.

What GitOps Doesn't Solve

Environment Provisioning

GitOps syncs what exists. It doesn't create environments on demand.

When a developer needs a staging environment for a new feature, someone still has to:

  • Create the namespace
  • Apply RBAC bindings for the right team
  • Provision the database (Postgres StatefulSet, PVC)
  • Set up Redis
  • Configure the ingress
  • Wire the secrets

ArgoCD will happily sync your app manifests into that environment once it exists. But the environment itself? That's a manual job. On most teams, it falls to a senior engineer — the same one handling incidents and architecture decisions.

According to IDC research, engineers spend 20-40% of their time on infrastructure tasks rather than product work. Environment provisioning is a significant chunk of that.

Developer Self-Service

ArgoCD's UI is ops-facing. It shows application sync status, resource trees, and deployment history. It's built for platform engineers who understand Kubernetes internals.

A junior developer who needs a staging environment to test their PR doesn't know what an ApplicationSet is. They shouldn't have to. GitOps doesn't abstract the complexity — it just moves it to a Git repository.

Self-service means a developer can request and get an environment without understanding the underlying infrastructure. GitOps is a prerequisite for self-service, not a substitute for it.

Cost Visibility

GitOps has no concept of cost. ArgoCD doesn't know or care that a namespace has been running idle for three weeks with a Postgres StatefulSet consuming $180/mo in compute and storage.

It syncs state. Idle state is still valid state.

Without an active cost layer, idle environments accumulate invisibly. Flexera's 2025 State of the Cloud report found that organizations waste up to 32% of their cloud budget on idle and overprovisioned resources. GitOps doesn't address this — it can't.

Idle Environment Cleanup

Related: ArgoCD syncs to desired state, but desired state doesn't expire. A staging environment created for a PR that merged six weeks ago is still "desired" as far as ArgoCD is concerned, because nobody updated the Git repository to remove it.

A single idle environment — Postgres, two API replicas, a load balancer — costs $180-250/month. Multiply by the number of forgotten staging branches in a typical team and you have a significant bill with no corresponding value.

Incident Diagnosis

When something breaks, GitOps tells you what changed in terms of deployments. It doesn't aggregate logs, analyze resource metrics, or suggest root causes.

kubectl describe pod and log diving are still your primary tools. GitOps doesn't change that.

The IDP Layer

An Internal Developer Platform sits above your GitOps tooling and fills these gaps:

  • Environment provisioning — declarative environment specs that create namespaces, RBAC, databases, and services automatically
  • Developer self-service — a UI (or CLI) that developers can use without Kubernetes knowledge
  • Cost visibility — per-namespace, per-team cost attribution with real numbers
  • Idle detection — AIOps that identifies abandoned environments and either flags or auto-removes them
  • Incident diagnosis — log aggregation and LLM-based root cause analysis

Crucially: an IDP doesn't replace GitOps. It uses GitOps. The environments an IDP provisions are managed by ArgoCD. The two tools operate at different layers.

How Fortem Works With ArgoCD

Fortem is a Kubernetes Operator that manages environment lifecycle. When a developer creates a FortemEnvironment resource, the Operator provisions the namespace, RBAC, and constituent services. ArgoCD then syncs the application manifests into that environment.

apiVersion: fortem.dev/v1alpha1
kind: FortemEnvironment
metadata:
  name: feature-payments-v2
  namespace: team-backend
spec:
  template: microservice-stack
  services:
    - name: payments-api
      image: registry.internal/payments:pr-442
    - name: postgres
      preset: postgres-15-small
    - name: redis
      preset: redis-7-ephemeral
  ttl: 72h

Enter fullscreen mode Exit fullscreen mode

ArgoCD watches the namespace Fortem creates and syncs the application config. When the TTL expires, Fortem tears down the environment — including notifying ArgoCD to remove the associated Application.

The separation is clean: Fortem owns the environment lifecycle, ArgoCD owns the deployment sync. Neither duplicates the other's job.

Key Takeaways

  • GitOps solves deployment consistency and drift detection. It doesn't solve environment provisioning, developer self-service, cost visibility, or idle cleanup.
  • The gap between "GitOps in place" and "developers can self-serve" is real and significant — it's where 20-40% of engineering time gets consumed.
  • An IDP operates at a different layer than GitOps. They're complementary, not competing.
  • If your platform team is still manually provisioning environments for developers, you have a self-service gap that GitOps alone won't close.

If this gap sounds familiar, Fortem is a self-hosted Kubernetes IDP that installs as a single Helm chart — no dedicated platform team required. Free tier available.