[HN Gopher] The True Cost of Kubernetes: People, Time and Produc...
___________________________________________________________________
The True Cost of Kubernetes: People, Time and Productivity
Author : edouardb
Score : 25 points
Date : 2021-12-30 15:59 UTC (7 hours ago)
(HTM) web link (www.koyeb.com)
(TXT) w3m dump (www.koyeb.com)
| ZeroDeth wrote:
| Great article. Thanks
| amscanne wrote:
| > For all these offerings [GKE, EKS, AKS], there are no automatic
| version updates or auto-recovery and you still need to pay for
| the computing resources like CPU, memory, and ephemeral storage
| that your worker pods consume.
|
| I don't know about the others, but GKE certainly has automatic
| updates (for masters and nodes). There's also auto-repair and a
| backup feature (is this "auto-recovery"?). GKE has also had
| autopilot for nearly a year, which bills based on Pod requests.
| PaulHoule wrote:
| I think they could get around the problem of requiring a third-
| party cluster manager if they really thought through the chicken-
| and-egg problems, and concentrated the circular dependencies into
| a very small core.
|
| I've had the experience of working through the design of a system
| over and over until circular dependencies can be tamed through a
| bootstrap procedure.
| icedchai wrote:
| These costs are highly inflated. Not sure why you need 4 people
| to operate a small 6 node cluster. From my own personal
| experience, one guy can do that part time. Your cluster has
| redundancy, so most problems can wait to be dealt with during
| regular hours.
| zufallsheld wrote:
| Why do I need four people for on call with a self hosted cluster
| but only one for a managed cluster? My application still needs
| the on call support. I think this cost estimate does not add up.
| yann_eu wrote:
| You're right, having only one person on call for a managed
| cluster doesn't make a lot of sense. We should probably have
| planned with at least 2 people for a managed cluster too to
| cover 24/7/365 operations.
|
| I think our thought process here is that developers are also
| involved in on call support for the service availability and
| the k8s cluster availability is mostly managed by the provider,
| but the cluster can still fail even if the control plane is
| managed.
| dijit wrote:
| Self managed cluster needs networking, some kind of
| persistent volume storage and the nodes themselves need to be
| somewhat maintained.
|
| I think you could get one person to be on-call for all those
| things, personally. But then I think that person should not
| be on-call for application support (IE; not the things
| running inside k8s, they would be the person the on-call
| application developer/administrator would call if they
| couldn't debug issues with networking, for instance).
| mkhnews wrote:
| What about persistent storage and access costs ? And couldn't
| that then over time increase managed k8s costs significantly ?
| yann_eu wrote:
| We focused on estimating the minimum/entry-level cost of
| Kubernetes here.
|
| If you have a data intensive service, it surely would add up,
| but it's not specific to Kubernetes. If you go with VMs or a
| Serverless deployment, you'll have to pay it too.
|
| If you're speaking about the storage and data transfers related
| to the Kubernetes control plane itself, I don't believe it
| represents a significant cost, even with a large cluster.
| yuppie_scum wrote:
| It's a lot easier now with EKS etc maturing. I think people will
| develop some more wrappers to make things easier.
|
| Regardless, the alternative mentioned (Serverless) comes with its
| own set of problems.
| renewiltord wrote:
| Meaningless analysis. If you're the kind of org where running a 6
| node cluster needs 4 people performing continuous ops, you're not
| going to succeed. Good luck.
___________________________________________________________________
(page generated 2021-12-30 23:01 UTC)