GitLab CICD Pipelines with AWS EC2 and S3

[@alexmoiseev5442]

Marty - Thank you! Concise and straight to the point!

[@Vnairtruered17]

Hi Marty, this was fantastic. I was looking for a straightforward way of using CI/CD in Gitlab, this video really helped. Thanks.

[@jenniferclark7094]

Really great Marty - simple, clear and helpful.

[@mslorraine101]

This was a great overview. Everything is tangible and easy to follow.

[@farodyne]

GREAT tutorial! You really are good at explaining things!

[@Learn_IT_with_Azizul]

very nice explanation. but, sound is so low... please amend it. Ty

[@michaelkolawole2311]

Awesome tutorial Marty. can you share the yml file?

[@wysefavor]

Hi Marty: Are the servers stage-express-api and prod-express-api the gitlab runners ? I want to know how gitlab authenticate with AWS, the role "s3-admin-access" was given to both servers for authentication right ? 9:53

[@pigrebanto]

why the repo is not shared?

[@fringefringe7282]

What you are doing is an anti-pattern. You should run the pipeline at least for unit and integration tests on merge requests to dev to have short feedback loop. What you are doing can cause a lot of headache on sprint end.

[@far-red]

putting app's config in s3 is logically sound, and i agree, very secure, only issue is what do you do if a. your dev team is made up of 4 people, b. new config keys are always created with most development this result in having a need to update the config file pretty regularly, what s3 does support version control, but its the devops guys who maintains the config (add, modify key) with every merge request, if manually need to remember to update... is manual, and easy to cause issue if forgotten, and diff config for diff environment, usually is known better by developers.. bit of an issue, hence i say, put all config in git (yep not ideal), need to find an alternative,... anyone in the same predicament as me?

[@amanbabbar1709]

We can use hashicorp vault, conjur key management for configuration

[@onatkorucu842]

i don't fully understand what you are saying but generally git hooks are used to state rules to make sure noone forgets stuff.

[@boh3607]

can we do this with terraform ?

[@nguyentuananh9163]

Thanks

[@ArbazKhan-vp5hn]

Can anyone share the gitlab yaml file for this

[@swadhinpanigrahi7497]

gitlab option is not visible in s3 bucket

[@evanandrews8858]

Hey Marty, you have no idea how long I've been searching for an example of this workflow. Specifically deploying to EC2 rather than ECS. So, thank you for that. Do you know why this might be the case? I just want to run a simple EC2 server, I don't know why I feel like everything is trying to shoehorn me into an ECS...

[@levinskipolish344]

It depends on what type of architecture you are aiming for. If you want a stateless machine to run your application at a scalable fashion, ECS is the way to go. It helps you to keep the same base code while expanding horizontally throughout a cluster of EC2s.

[@amanbabbar1709]

Marty do you have repo for this tutorial

[@noherczeg]

Never ever deploy to PROD from a pipeline directly.

[@Ничтожество-и5ш]

Why?

[@piotrfilochowski1646]

@@Ничтожество-и5ш Because pipeline deploy may randomly fail, it happens like 1 per 8 deploys for me. And you want to have PROD deploy stable, testable, predictable as it can be. So I'm deploying to PRE-PROD environment, and when it's tested, I'm promoting it to PROD environment. (I'm taking an already build jar file from one environment to another one).

[@479shan]

Speak louder