Eureka! Finally got over the last hurdle - trying to figure out how to get a generated zip file over to S3 from gitlab CI. Gitlab CI doesn't have zip by default, and no combination of tar options seems to make S3->CodeUpdate happy.

Thanks, I eventually slogged through it.

So my flow is - merge to develop branch in gitlab.com, which triggers a gitlab CI/CD build and deploy to S3, which is either the end of the line for front-end code, or a specific file appearing or updating in an S3 bucket triggers a code update to a lambda function. Fun stuff.

Next step is cloud formation templates.

Asking for a friend - if you're super-happy about all aspects of your current job and your company is also a great brand name that attracts recruiters, etc, and you're being recruited into a risky situation - say a company that is a lesser name in the industry and you're not as sure about how happy you'd be at the new company, etc - what would be roughly the % increase in comp that would justify taking the risk? Let's say that the comp increase also roughly corresponds to the increase in seniority. Obviously there's no right answer but curious to hear everyone's take.

50% for sure. 20% probably not.

Keep in mind that you will never rise up the ranks w/o hopping stones like this. Unless you're on the fast track. And if you aren't sure if you're on the fast track or not - you aren't.

Fast track is gonna be some guy who went to the same Stanford frat as your boss's boss. Of course they don't go around advertising that. But they know.

But if you're fine with just pulling in a nice salary + stock for a long time - then that doesn't matter.

A buddy of mine had a nice career at Sony - they even sent him to Tokyo to work for 3 years to learn the business. But at some point he realized he wasn't gonna ever rise too high up the ranks - so he jumped ship as a VP of a much smaller company.

I'd like to say that worked out well for him. But it ended after a few years. He's currently underemployed with some scattered side stuff - and getting frustrated looking for work. He didn't want to hear about my awful ordeal of 1 month of job interviews - lol.

Basically don't be pushing 50 (which I know you aren't) and out there w/o a life raft imo.

I think the main question you need to ask your friend, is "where do you want to be in 5 years", and then figure out which job will get you there. I'd need 30-40% raise and a better title.

So one of the two staff engineer just told me he’s going to be taking a two month sabbicatical. He said he’s going to spend some time thinking about what he wants to do afterwards. There’s a good chance that he won’t come back.

Which is really unfortunate since I have been learning a lot from him. And he also has been playing a lot of the product manager role since we have no manager for our team yet.

I’m kind of worried since we have a huge product release we want to do in Q4 and I’m one of two guys who have a good understanding of it. We have a senior engineer who just joined two months ago, that is supposed to step up and help lead but my confidence of his abilities is quite low. So far he hasn’t shown much technical prowess nor has he has done a good job understanding of our business domain. The other staff engineer ( the one who said I was trolling) has a good grasp of the business domain but will be stepping away from getting his hands dirty coding.

So it will be interesting to see what will happen in the next couple months. Luckily we have a new grad who is a returning intern who has been freaking killing it lately. It will be refreshing to have him along my side leveling up the team with Scala.

I think age and maybe networth plays into the question, not just percentage increase in comp.. Someone with a family and kids isn’t going to want to jump on this opportunity. It’s a high risk - high reward situation.

you'll always have that company on your resume, and it's likely you can go back if it doesn't work out. if you think you're capable of bigger things, then you have to be open to the opportunity. if you think you're at about the right level in terms of salary and position, and you're still learning, then stay put.

Second this. To grow you should strive to be antifragile.

Are the personal images in docker hub vetted in any way? Like can I trust the one with aws cli pre-installed isn't scraping my AWS credentials and sending them off somewhere?

I doubt it.

I think they always include a Dockerfile though, so you can recreate it yourself.

Ok so this looks pretty safe: https://github.com/sekka1/aws-cli-do...master/1.15.47

Can I trust that's the image uploaded to docker hub? https://hub.docker.com/r/garland/aws-cli-docker/

IE - does docker generate the image or do you just upload something and link out to a dockerfile that you say generated it?

https://ropenscilabs.github.io/r-doc...Dockerhub.html Hmmm - doesn't seem like there's any check on docker's end to make sure the image matches the dockerfile?

I’ve only been working sporadically with docker for a few months but i would take caution in downloaded images, which it looks like you are doing

The other option is to install aws cli on each build. Which slows things down considerably. I guess I should just upload my own docker image.

We basically make custom base images that include stuff that nearly every image needs, and then inherit from that. So we'll start with the ubuntu version we want, add whatever packages and libraries, etc. This is a good time/place to put our own python packages that nearly every project uses at least some part of.

We don't use dockerhub to manage our own images - we use an "artifactory", in our case we use both jfrog and AWS's artifactory. (I don't know why we use both, but it's 2 different teams. One probably predates the other and we haven't totally switched)

I'm using gitlab hosted CI - so I'm not sure I can pull from a custom repository.

I just closed 12 tabs I opened yesterday trying to figure out how to create a zip file with tar. Still never could figure out how to do it. AWS can’t seem to unzip anything created with tar - at least for lambda code update.

I've never used tar without passing like 4 options -zkcd like why would you have a command that everyone has to use every option on every time? Have they heard of defaults?

Don't remember where I got this from but it's helpful for unzipping. I should probably make something for compressing too.

That looks nice.

Haha that seems like overkill. I just have all the commands mostly memorized by now.

I don’t really like scripting in that way because it encourages me to forget how things actually work when I am writing a for real script and need to use a command.

One of the things I’m concerned about for my development as a devops person is my familiarity with shell scripting and almost no knowledge of python. From what I understand python is the main scripting language and shell is kind of going the way of the dinosaur. I could be wrong, but that’s the impression I’ve come under. We only use shell scripts.

He is doing well enough that there's no real downside financial risk associated with either choice. He'll be fine either way. The risk is more about known vs unknown and a place where he is certain to be happy vs a place where he may not be, as opposed to taking some kind of big career risk.

Is this truly the minimum you'd need? What if there's no immediate prospect of a raise in your current job? What about if you're pretty sure your comp is going up ~15% in a year?

I'm curious because lots of people choose companies that are both objectively worse brand names and pay less (those are usually correlated) so I can't imagine the brand name premium is worth much beyond the fact that brand name companies pay better in the first place.

Well "right level" is market-dependent and "learning" doesn't preclude a higher rate of learning elsewhere. I'm just curious how you would price this situation. Or if not %, how much more per year would you need to jump in this situation?

What about 25%?

Whoa, what was his role? Was he never able to find anything after that VP role or were there other gigs that also didn't work out afterwards.