Cloud, Ops & Security · Frontier
Kubernetes
A tool for orchestrating a large fleet of containers: if one disappears it starts a replacement, and when traffic rises it adds more, all without a human doing it by hand. You only need it once the scale outgrows manual management.
TL;DRIn 30 seconds
- A tool for orchestrating large numbers of containers: replace the dead ones, add more when traffic rises.
- Once containers multiply, managing them by logging into machines one by one goes wrong — hand it over and it arranges everything.
- Small projects don't need it; it earns its keep when the scale outgrows manual care.
OverviewIn detail
Kubernetes is often written K8s, because exactly eight letters sit between the K and the s. It is a container orchestration system, and what it manages is containers — the self-contained units that package a program together with the environment it needs in order to run. Once there are many containers, a string of questions has to be answered: which machine should each one live on, should a crashed container be replaced, should more be started when traffic rises. All of that is left to Kubernetes to arrange automatically.
What it solves is a scale problem. With dozens or hundreds of containers, logging into each machine by hand to start, check and restart them is bound to go wrong and is more than anyone can keep up with. Kubernetes continuously steers the overall state of the machines towards the target you declared: say you want three copies, and it keeps three running, replacing one the moment it dies.
It is not the same thing as Docker, and the two can't stand in for each other. Docker solves how to package and run a program on a single machine; Kubernetes solves how hundreds of such programs spread across dozens of machines are coordinated, scheduled and scaled. The former packages and starts, the latter orchestrates; they are usually used together, one handling the local, the other the whole.
For an individual it usually stays far away. Running one or two small projects doesn't call for anything this heavy. You are most likely to meet it in enterprise deployment plans or in descriptions of agent platforms — talk of multiple clusters, edge nodes and autoscaling is it orchestrating across a large fleet of machines. Getting hands-on has a steep learning curve and no small operational cost; it isn't something you spin up casually.
MythsCommon misunderstandings
- ✕Thinking it's the same thing as Docker ✓Docker handles packaging and starting on a single machine; Kubernetes handles how hundreds of containers are scheduled across many
- ✕Thinking adopting it guarantees stability ✓It steers towards the target you set — set the wrong target and it faithfully runs in the wrong direction
- ✕Thinking small projects should use it too ✓One or two little programs have no use for it, and installing and maintaining it is real work
In practiceWhere you will meet it
- Meeting it in your company's deployment plan, where it places hundreds or thousands of containers
- Traffic spiking overnight and it adding instances by itself, with nobody working late
- Running nothing but one small website, in which case you'll rarely touch it
SummaryThe key points
- An orchestration system for managing large numbers of containers
- Replaces dead containers and adds copies when traffic grows
- A different job from Docker: it schedules, Docker packages and runs
- Individual projects rarely need it; enterprise deployments do
FAQ
What does Kubernetes mean?
A tool for orchestrating large numbers of containers: replace the dead ones, add more when traffic rises.Once containers multiply, managing them by logging into machines one by one goes wrong — hand it over and it arranges everything.Small projects don't need it; it earns its keep when the scale outgrows manual care.
Where will I run into Kubernetes?
Meeting it in your company's deployment plan, where it places hundreds or thousands of containers;Traffic spiking overnight and it adding instances by itself, with nobody working late;Running nothing but one small website, in which case you'll rarely touch it。
What do people most often get wrong about Kubernetes?
Thinking it's the same thing as Docker — but Docker handles packaging and starting on a single machine; Kubernetes handles how hundreds of containers are scheduled across many;Thinking adopting it guarantees stability — but It steers towards the target you set — set the wrong target and it faithfully runs in the wrong direction。
What are the key points about Kubernetes?
An orchestration system for managing large numbers of containers;Replaces dead containers and adds copies when traffic grows;A different job from Docker: it schedules, Docker packages and runs;Individual projects rarely need it; enterprise deployments do。
Common pitfalls Before adopting Kubernetes, confirm the scale really calls for it — on a small project the burden usually outweighs the benefit.
Also written asYou may also see it written as
These are all the same concept — the same thing, however it is written in a skill description. Do not mistake different spellings for different ideas.
Entry 169 of 186 · All entries
Keep readingRelated reading
DevOps(开发运维一体化)
让写代码的和管服务器的人共用一套流程和工具,把测试、打包、上线交给机器自动完成,改完代码很快就能安全发布。
镜像(Image,软件打包模板)
把软件和它运行需要的环境一起打包好的模板,取出来在任意一台机器上都能跑出一样的结果。
HTTPS(加密网页协议)
给浏览器和网站之间的通信上锁加密,网址前面那个小锁就是它;路上有人偷看,也只能看到一串乱码。
漏洞(安全漏洞)
软件里一处能被坏人利用的缺陷,能让对方做本不该被允许的事,比如偷数据、冒充管理员或把程序弄崩。
会话(Session)
服务器用来记住这一段连续互动还是同一个人的方式;从你登录到退出之间的操作,被串成一整条记录。
Linux(操作系统)
免费开源的电脑操作系统,全球大多数服务器都在跑它;通常没有图形界面,但稳定、省资源、能长期不关机。
Where to go nextExplore more
Back to the skills themselves
Most of the concepts covered here come from skills that install into your AI client with a single command.