[02:20] <supremekai> Hey guys, not sure if this is the best channel to ask this, but I would like to get a cloud provider VARIABLE RAM/CPU instance to mount a ubuntu-server. Where can I get it? I am using a VPS from VULTR which has a FIXED RAM/CPU, but I am not being able to find a VPS with variable RAM/CPU usage, could you provide me some reference? Like Google Cloud?
[02:24] <sarnold> my irssi aws instance is a lightsail instance; memory is fixed :( -- all the docs suggest that it's easy enough to change the memory of normal aws ec2 instances, but lightsale instances are set in stone
[02:25] <sarnold> I don't know about the CPU -- instance types feel very fixed, but perhaps the vcpu count is more flexible? I'm not sure :( I bet you could test a few things on the free tiers of aws, azure, gce, oracle, and figure out what you like and dislike without costing anything
[02:48] <supremekai> sarnold, yup :\ I was looking for something that would allow some RAM/CPU peaks that my backend is experiencing
[02:49] <supremekai> but I can't seem to find any cloud instance which provides this flexibility
[02:49] <supremekai> I've always used fixed-size CPU/RAM machines
[02:49] <supremekai> but with the advent of google cloud, aws, etc.. we could have some kind of service like that available
[02:49] <supremekai> the issue is that they have so much products that I cannot find it specifically
[02:50] <sarnold> supremekai: isn't that the truth :( I hate the aws docs so much
[03:05] <keithzg> Yeahhh there's some longstanding possible improvements I could make to the backup setup at my work but so much of it would have to do with reading deeper on AWS and I always find a way to shuffle that task lower in the queue
[11:44] <tomreyn> FWIW, on AWS you have the AMD CPU instace types (ending in 'a' as in AMD, such as "T3a") which are burstable, i.e. allowing for CPU overcommit at an increased price (can be disabled): https://aws.amazon.com/ec2/instance-types/#Burstable_Performance_Instances
[11:46] <tomreyn> but their general answer to increasing + lowering demand is autoscaling - adding or removing instances based on demand - which requires a high availability software / system architecture
[16:08] <patdk-lap> hate aws docs? they are so good, compared to most everyone else
[16:09] <patdk-lap> if you need more flexable instance sizing, google is much more flexable, but I don't know any that will do dynamic sizing
[16:10] <patdk-lap> personally I'm using k8s on aws using spot instances and let it autoscale the cluster up and down
[19:15] <sarnold> heh yeah it's entirely possible that the aws docs are vastly better than others; I just know that I can't pay attention to them for any length of time.