How can i learn aws?? Any resource?

Please keep uploading Dax

@10:37 FYI, it does not simulate database engine, it actually is the engine that they customize it.

Hello, I was always afraid of using spot instances on production. I am still scared of doing it. This additional moving element can break, and I am too scared to apply it. I wish you luck in using spot instances for long-running processes. For late cron jobs and some machine learning stuff, it is perfect.

If you go on Bare Metal Savings you'll find that AWS becomes too expensive much too quick compared to outright owning today's _insane_ hardware. If you need the flexibility of the cloud, deploy cloud agnostic software to on-prem then offload to cloud on peaks, you'll save a lot of money.

9:01 I believe Fargate is Serverless Ec2, and On-Demand is a concept for normal Ec2 instances.

Is there a way to do the inverse of item 4, where you would be able to reference a Resource inside a non-sst piece of code?

0:35 Get wrecked kid.

Reserved instances aren’t a legacy concept?

this channel is quickly becoming a gold mine

Is sst as powerful when used with non AWS providers like GCP ? You do most of you examples with AWS, is there a reason for that ?

Any chance `capacity` will come to Cluster Tasks?

audio is off

You shouldn't pay for an on demand DB server until you have at least 10 customers.

SST content is where it’s at!

Another great feature Dax!

Remove the 30 RDS instances and 100 NAT gateways once you are done developing for the day or sell your personal belongings if you forget 😂😅😂😅😂😅 SST Rocks, thank you guys so much.

When did dax grow hair?

Heres an edge case, curious what people would do: I made an SST Linkable with Cogntio pool and a cognito pool client. Gave it all cognito-idp permissions. But when I tried to use it things were complaining that the pool client wasnt linked even though I had the ID in my linkable

we got sst paid plan before gta6? lets f go

but if you only want to apply the config of the Linkable.wrap method to a one instance of the component only, how can you do that?

Please start addings links to the video description. Not a big deal but it helps ☺

Please continue doing these!

I'm making a web application in uni right now, and .env is PERFECT for what we're doing. It's easy to set up, the only thing we use it for are throwaway test accounts on Atlas and stuff. I agree if the local environment has access to paid services, but that's not what we're using it for anyway. We just have to be careful not to put any sensitive data in our local environment, that's all.

man you look good and thanks

thanks jay, damn you look good!

i wanted to love dynamodb but thru appsync, oh the pain of a blind truffle pig aka noob developer. supabase is so much easier.

go, php & js are the best! also bash

is there a 'go fuck yourself' command?

Thanks for the comparison. At least in Spain, it seems that RDS has a minimal cost, lower than Aurora, but it does have it. That's why I chose DynamoDB, because it really is free unless your project starts working with real data and users. You can develop on it for months at no cost. It was difficult for me to fully understand DynamoDB and all its implications, but after two months I'm starting to feel better about using it (30 years of RDBMS is a long time...).

A little confused. You can spin up an aurora instance inside RDS. Are you talking about Aurora serverless?

Great video!

Would appreciate so much adding the new DSQL to this!

These videos have been great, stuff I've wondered about and quick breakdowns of web service options and tradeoffs.

Neptune?

static linking is such a great feature, with the linux syscall API being so stable you can just take an executable and run it anywhere and it just works.

Just discovered your channel, thanks!!

surprised no mentions of dynamo and dsql

i learn from your videos alot!!

This is a good thing, but the selling point “we created something new and a bit unusual” said at the beginning is misleading. Literally every company I’ve ever worked in does this

Thx for the video! I believe it would also interesting to mention how applications manage connections. How do we know if the client is using the replica for reads and the primary one for writes? Will we instantiate more than one client? Do we have to properly configure the client? Do we have just a connection string and everything is handled by AWS under the hood? Those kind of things... P.S: even I don't really like the "lock-in" argument since I don't know anyone who moved all its infra from vendor A to vendor B (or even got vendor rekt - DHH mentioned), would the Aurora usage prevent data export/import to another vendor and database?

What about DynamoDB? And how it scales etc compared to these

0:03 every single week 0:04 every single week

no dynamo?

cool video

Great video man! Simple but informative explanations

Very useful video, thank you!

DynamoDB is the winner for me.

I just reliving the incident recovery training few months back with this video 😂 PITR is tricky but saves your butt.

Aurora serverless v2 showed pretty bad in our case because it always was adjusting max_wal_senders back and forth during spikes and restarting the db. There was no option to freeze that so we just moved to a regular aurora