Gitops Considered Harmful for MVP
Posted3 months agoActive3 months ago
knockdata.comTechstory
calmnegative
Debate
40/100
GitopsMVPDevops
Key topics
Gitops
MVP
Devops
The article argues that GitOps may not be suitable for Minimum Viable Products (MVPs) due to its complexity, and the discussion revolves around the trade-offs between GitOps and other approaches.
Snapshot generated from the HN discussion
Discussion Activity
Active discussionFirst comment
2m
Peak period
13
0-12h
Avg / period
7
Comment distribution14 data points
Loading chart...
Based on 14 loaded comments
Key moments
- 01Story posted
Sep 26, 2025 at 3:41 AM EDT
3 months ago
Step 01 - 02First comment
Sep 26, 2025 at 3:44 AM EDT
2m after posting
Step 02 - 03Peak activity
13 comments in 0-12h
Hottest window of the conversation
Step 03 - 04Latest activity
Oct 2, 2025 at 3:42 AM EDT
3 months ago
Step 04
Generating AI Summary...
Analyzing up to 500 comments to identify key contributors and discussion patterns
ID: 45383862Type: storyLast synced: 11/20/2025, 4:23:22 PM
Want the full context?
Jump to the original sources
Read the primary article or dive into the live Hacker News thread when you're ready.
> At its core, GitOps is simple. You write down how you want the system to look in code and a bot makes the world match that. Your infrastructure lives in Git. You deploy by committing. You roll back by reverting. The robots take it from there.
For version 1/MVP work, you absolutely shouldn’t bother with this. It’s a complete waste of resources when you should be focusing on growth or launching the product. Compared to doing it by hand, it’s slower, clumsier, and just another layer of complexity your team has to deal with.
On the other hand, for long-running, stable systems, it’s awesome! We know exactly who rolled out a change and when. From the commit messages, we know why the change happened—even years later. We also make a point of adding Jira (Hawk Tuah) ticket numbers so we can track the details more easily. And if something goes wrong, it’s simple to roll back to an older version.
This approach is perfect for large, long-term maintenance systems—but poison for a brand-new project.
In theory, GitOps is neutral. A robot pulls from Git and makes reality match. Everyone gets to review, and every change is versioned. Feels fair. Right?
But in practice, GitOps introduces a very specific kind of power dynamic: the gatekeeper pattern.
Most of the time, it’s the infra or platform team that sets up GitOps. They define the rules—how environments are structured, how approvals work, which tools are allowed. And once that system is live, every change has to go through them.
It sounds like collaboration. In reality, it’s almost always a one-way review.
A backend developer wants to change a config file. They need a review from someone on the platform team. A frontend dev wants to bump a service version. They open a PR. They wait. A product engineer wants to expose a new route for testing. Same story. PR. Wait. Fix a nit. Wait again.
But it doesn’t go the other way. It almost never goes the other direction.
Infra changes things, merges to main, the bot deploys it. No one outside the infra team is reviewing their changes. No one’s stopping their PRs with a comment. They own the system, and everyone else is a guest.
That’s not collaboration. That’s control. """
Obviously, the level of auditing and reviewing for infrastructure changes in a Prod environment make no sense for a Sandbox environment, and there’s nothing in GitOps that implies these need to be the same.
Ideally at every phase of development, you have very legible infrastructure that can be shared and iterated on by a team. The CI pipelines backing this should offer rapid turnaround times, and things should be easy to test.
All things which the general GitOps concept still works in tandem with.
The PR review cycle is not a good fit for MVP (and to internal tools, etc.) . I think, GitOps still can be used if
* You have a sandbox / test environment * Allow the developers to push directly (no PR) to the branch used for that test environment
So for me it's not about GitOps (having the description of the deployment in git) it's a about the git workflow for these cases (PR review cycle vs pushing directly to branch).