Products · Product
Kubernetes
Raised as a usability problem by 238 people on Hacker News over 12 years (2015–2026), in 186 threads.
No issue was raised by at least 3 people in 3 threads this year.
- 1
- 2
- 3Writing configuration in YAML is awkwardUnreviewed
- 4
- 1
- 2
- 3Writing configuration in YAML is awkwardUnreviewed
- 4
- 5
- 6
- 7
- 8
- 9
- 10
Every issue, all time, most people first
Cluster setup is nearly impossible and configuration is hard to discoverUnreviewed
20152026- Too easy to set up clusters without understanding them, so problems are hard to troubleshoot2024-11-02
- Specific configuration keys are hard to find, so searching wastes hours2022-07-03
- Too complicated to learn; couldn't manually set up a three-node control cluster after two weeks2022-06-14
YAML errors point to line 1, leaving the cause unknownUnreviewed
20152026- Cryptic YAML errors pointing to line 1 leave the user guessing what went wrong2025-08-31
- Confusing commands and YAML files that constantly break and need copy-pasted fixes2024-11-27
- Deeply nested YAML is hard to read and write, with no schema checking2024-01-05
Writing configuration in YAML is awkwardUnreviewed
20152026- YAML templating with Helm becomes a horror to work with2026-06-17
- YAML and Helm templates are terrible to write and configure2026-06-16
- Too many overlapping tools just to template and deploy YAML2024-03-27
Documentation is inconsistent and hard to find accurate information inUnreviewed
20152026- Docs lack versioning, links are dead, and too much text with few examples2024-12-12
- Single-page docs are hard to search; three patch types cause integration effort2021-06-22
- Documentation is confusing and scattered across too many places2019-01-20
Abstraction layers are hard to understand after time awayUnreviewed
20152026- Layers of abstraction make it hard to understand what is actually running2024-11-26
- Manifests are a complicated mess of layers that hide complexity, making it hard to work with2023-09-22
- Multiple overcomplicated layers and confusing documentation make problems impossible to figure out2021-02-11
Docs don't state minimum setup for self-hosted clustersUnreviewed
20152026- Docs don't clearly say the minimum setup needed for self-hosted clusters2024-11-28
- Hard to set up and manage clusters without cloud providers, making self-hosting painful2023-03-23
- Documentation doesn't explain how to make a working cluster or expose services2018-03-28
Too configurable with too few abstractions, adding complexityUnreviewed
20152026- Too many competing ways to do things lets developers add complexity they can't handle2025-01-18
- Too configurable with not enough abstractions, so users face complexity2024-09-01
- Too many added abstractions and extensive API make it complicated to use2022-02-01
One-line YAML fixes, networking, permissions and slow setup are gotchasUnreviewed
20152026- Constantly changing YAML and beta API versions break things after cluster upgrades2024-06-26
- Running code requires hand-crafted YAML and heavy cluster configuration2021-09-20
- Many gotchas: one-line YAML fixes, networking issues, hard permissions, slow cluster setup2019-10-22
kubeadm needs manual setup of monitoring, logging and authUnreviewed
20152026- Kubeadm clusters need lots of manual setup of monitoring, logging, and auth tools2025-06-20
- Easy to make persistence errors when running databases on Kubernetes, and backups and recovery need more manual work2018-12-31
- Hard to automate rollouts with kubectl without pain2016-09-27
Cannot diagnose DNS failures across many interacting componentsUnreviewed
20152026- Layers of abstraction hide DNS problems, making failures very hard to diagnose2022-08-10
- Too many confusing network options; DNS problems hard to diagnose and fix2022-06-27
- Hard to understand what runs under the hood; can't diagnose DNS failures with many interacting components2019-09-20
Unreviewed issues were grouped by machine: Claude Haiku labelled each comment as a usability complaint, named what it's about and summed up the problem in a line, and a small model grouped similar problems. Nobody has checked them by hand. Expect mistakes.
The newest 50 complaints
- Cannot easily check one service's status across many clusters at once2026-08-02
- Hard to use, debugging is painful, nostalgia for older simpler WebSphere2026-07-06
- Unnecessary complexity and many failure modes make services hard to debug and operate2026-06-25
- YAML templating with Helm becomes a horror to work with2026-06-17
- YAML and Helm templates are terrible to write and configure2026-06-16
- Settings exist but nobody documents how to change them or their effects2026-04-27
- Kubernetes is crazy difficult to set up and manage2026-04-23
- Hard to navigate nested sub-commands in verbose help output2026-02-20
- Memorizing and retyping CLI commands is slow and typo-prone during debugging2026-01-05
- Getting secure secrets, networking and storage working needs lots of extra add-ons and work2025-11-14
- Hard to keep up with frequent changes to the platform2025-11-14
- Hard to learn and to build or maintain on bare metal2025-11-12
- Cryptic YAML errors pointing to line 1 leave the user guessing what went wrong2025-08-31
- Same name 'nginx' used for four different things, so it's hard to tell what does what2025-07-14
- Too much configuration to maintain, duplicated with local compose files2025-06-23
- Can't easily get logs of a custom resource; must dig out its controller pod manually2025-06-21
- Overly complex, full of pitfalls, needs heavy documentation and training to use2025-06-20
- Kubeadm clusters need lots of manual setup of monitoring, logging, and auth tools2025-06-20
- Running Kubernetes outside cloud providers is a disaster waiting to happen without expert skills2025-06-19
- Finds Kubernetes API and UX very hard to use compared to Docker Compose2025-05-07
- Cluster stalls, fails to reach consensus with even node counts, and needs manual config file edits2025-04-05
- Lightweight Kubernetes distributions still cause huge CPU spikes from unwanted metrics components2025-04-01
- ConfigMap changes silently not picked up with subPath, causing hours of confusing debugging2025-03-11
- Annotation-based config is magic, undiscoverable outside docs, and unvalidated2025-03-10
- Setup requires many competing tools and mechanisms, making it far more work than simple autoscaling2025-01-18
- Too many competing ways to do things lets developers add complexity they can't handle2025-01-18
- Terrible first-time experience; hard for newcomers to get started2025-01-01
- Watch connections can lose changes or fail to reconnect, forcing developers to handle stale resources2024-12-16
- Docs lack versioning, links are dead, and too much text with few examples2024-12-12
- Descriptions hidden in collapsed nodes, so Ctrl-F fails and users click the wrong node2024-12-12
- Type definitions scroll away into a massive HTML page, losing spatial context and needing prior knowledge2024-12-12
- Setting up Kubernetes for personal use is hard, and deploying services requires large amounts of YAML2024-12-05
- Helm's AWS-like load balancing features are hard to understand without prior networking knowledge, leading to bad setups2024-12-02
- Kubernetes, ECS and Fargate are overly complex Rube Goldberg machines that need a team to decipher2024-11-28
- Docs don't clearly say the minimum setup needed for self-hosted clusters2024-11-28
- People copy large YAML blocks without understanding them and hope they work2024-11-27
- Confusing commands and YAML files that constantly break and need copy-pasted fixes2024-11-27
- No simple obvious setup; many manual steps needed to get basic functionality2024-11-27
- Newcomers can't understand the magic incantations in starter examples and can't navigate a running cluster2024-11-27
- Took a week to set up basics; documentation is hundreds of pages and poor2024-11-27
- Layers of abstraction make it hard to understand what is actually running2024-11-26
- Setting up job systems is a configuration nightmare that wasted weeks2024-11-05
- Too easy to set up clusters without understanding them, so problems are hard to troubleshoot2024-11-02
- Setup is far more complex than Heroku-style platforms2024-09-09
- Too configurable with not enough abstractions, so users face complexity2024-09-01
- Hard to understand what the system is doing; a random annotation breaks assumptions2024-08-23
- Controllers make resources behave differently than documentation says, making problems exhausting to trace2024-08-09
- Interface feels sluggish and slow to respond compared to older desktop apps2024-08-09
- Leaks implementation details and pushes too many config options onto developers2024-06-29
- Constantly changing YAML and beta API versions break things after cluster upgrades2024-06-26