Looking for an autoscaling container service
This request is perhaps a bit against the grain of this forum, which specializes in servers, but I thought I'd ask anyway.
So, I'm looking to host a container-based application, which I expect would have a spiky traffic pattern, and therefore I'm looking for a service which can automatically scale up containers based on the volume of HTTP requests, or alternatively, CPU and memory usage.
Typically I'd use HorizontalPodAutoscaler in Kubernetes, but I'd rather leave the sysadmin tasks to a service provider for this project.
The service can be in any location, as long as there's an AWS or GCP region nearby with < 10ms latency.
I've already excluded AWS Lambda and Google Cloud Run from the list - the containers need to be active for some time once a request has been handled, as I need to perform asynchronous processes in the background once a request has been handled. AWS's own container service, Elastic Container Service also seems to be a no-go since they charge $16 for a tiny container with half a vCPU and 512 MB RAM, and I think there may be cheaper services out there.
Comments
What kind of application you are trying to host?We can provide customized service to you billed based on the container usage if you require.You can message me if you want.We provide service in Hong Kong region which have both gcp and aws available nearby within 10ms
https://lowendtalk.com/discussion/186108/cloudjiffy-paas-free-14-days-trial-and-get-10-bonus-on-setting-up-a-demo-call#latest
https://cloudjiffy.com/
@leapswitch
i think they have autoscale and can give $10 if you signup for demo.
Fly.io has autoscaling in their documentation, I've not used it as I manage the resources required directly but I assume it probably works well as the rest of their platform does
https://fly.io/docs/apps/legacy-scaling/#autoscaling
Thanks for the recommendations folks.
@leapswitch have you improved your reliability meaningfully in the last 1/2 years? I used to be your customer, but had issues with containers randomly becoming unreachable, with support responding hours later with "seems like the issue is not happening any more," or the famous incident where you introduced SSH high availability by having different SSH host keys on different IPs (:facepalm:).
Good stuff, I'll take a look.
Hm...isn't this, what this "Jelastic stack" is meant for?
I remember playing with i.e. the solution from Aruba some time ago: https://www.arubacloud.com/jelastic-cloud-paas/jelastic-cloud.aspx
Worked well, but I never got the hang of it.
You could use google cloud run. Or find another jelastic provider. For I would just use vercel or cloudflare page, and call it a day
This is a containerized application that needs to do background processing after handling a request, so that excludes Google Cloud Run, as I mentioned above.
I'm probably going with fly.io, too many troubles with leapswitch and they followed up via PM but never bothered to reply.
FTR, I did receive a reply shortly after.
@fart Not sure if you taken a look at these or not:
https://caprover.com/
https://dokku.com/
vercel and cloudflare support background processes and cron job.
Also
https://www.virtuozzo.com/application-platform-partners/
They all are a jelastic provider much like cloudjiffy.