|
i'm new to aws and ec2, so forgive me if this is trivial. but is it possible to create a ubuntu instance somewhere in aws that's essentially a remote desktop? i'd love to be able to log on to this gui-ed instance from any browser or computer, and just do some dev work on there, which would entail remoting in to other servers and what not.
|
#
¿
Feb 2, 2019 19:50
|
|
|
|
| # ¿ May 12, 2024 01:17 |
|
|
i'm not the most experienced with aws, so forgive me if this is simple, but a friend is having a weird issue. basically, they have the two chains setup that goes like: dev chain: route 53 -> api gateway (cname alias) -> api gateway (custom domain) -> api gateway -> api gateway stage (dev) -> api gateway "api" (dev) -> elastic beanstalk (node.js) -> snowflake prod chain: route 53 -> api gateway (cname alias) -> api gateway (custom domain) -> api gateway -> api gateway stage (prod) -> api gateway "api" (prod) -> elastic beanstalk (node.js) -> snowflake but for some reason, the prod chain still ends up with requests calling dev. yep, he sees those requests in the elastic beanstalk logs, as well as in snowflake. he did some research in route 53 and the api gateway, and he believes the issue is one of the bolded links in the chains. he even redid those believed-to-be-wrong parts, and yet the prod chain still ends up with requests calling dev. like i said, i'm not an aws expert, or even all that good at networking, so i'm curious: what do you guys think? where should i look first? ever seen anything like this? my instinct is there's something overriding those portions of the api gateway and directing that one flow into dev, but i'm not sure what that'd be. or even if that's a smart instinct. in any case, my friend is worse than i am at all this stuff, so it's possible he got something very fundamental wrong somewhere. i'm not sure where to start, except for browsing around the api gateway. what he did show me there did look sensible and correct to my intermediary eyes, though. thanks for your input! and i can provide further details if that'd help.
|
|
#
¿
Mar 1, 2019 04:21
|
|
|
found it. looks like dude’s proxy endpoint got hosed
|
|
#
¿
Mar 1, 2019 22:48
|
|
|
is it possible to enable single sign-on for certain apps in an account and not the whole account? like, i want to know if it's possible to just have sso for quicksight and not, like, cloudformation or ec2 or anything like that. ultimately, what i'm trying to do is give some users in our network access to some dashbaords for testing. sso seems like it would be way easier rather than having to build and deploy accounts. at the same time, i don't want them to have sso access to the rest of the infrastructure. fwiw, i've researched it as much as my mind can possibly do. i'm not a network admin and i'm completely stupid about anything involving aws. i found this: https://docs.aws.amazon.com/singlesignon/latest/userguide/manage-your-applications.html which seems to imply it's possible if we move from iam to identity center? i don't know, aws is way too big and confusing for me to understand.
|
|
#
¿
Nov 28, 2022 20:17
|
|
|
i think what my boss wants me to do is figure out sso. basically, we want everyone in the product team and data team to be able to log on to the account, with the data team having complete access to quicksight and redshift and the product team having limited access to quicksight. like, the product team should only be able to view the dashboards. they shouldn't be anywhere near redshift either. at least that's how i understand it at the moment. and if i'm understanding both of your posts correctly, we'd basically just enable sso, then restrict the services using policies via iam and its identity center?
|
|
#
¿
Nov 28, 2022 20:41
|
|
|
thanks for all the help. seems like it's readily doable, just have to make, what are effectively, policy groups.
|
|
#
¿
Nov 28, 2022 22:22
|
|
|
so one of the dashboards i was tasked with creating in quicksight was supposed to consist of lots of scatter plots with regressions and what not. turns out, quicksight can't do that. HOW THE HELL DO YOU PROVIDE SCATTER PLOT VISUALS AND NOT ACCOMPANY THEM WITH REGRESSION OPTIONS?!?!?! i am utterly loving baffled by this. the biggest reason i can think of for using a scatter plot is to plot a regression on it. hell, i even talked to our rep who connected me with one of the lead engineers for qs. you just...can't do this. at all. apparently i have to either a) calculate the (x,y) for whatever regression i want in a table and then plot that on top of the scatter plot via reference lines, i think he said, (utterly ridiculous) or b) do all of this through sagemaker, which i know nothing about. loving gently caress. unreal.
|
|
#
¿
Jan 11, 2023 22:12
|
|
|
yea, even before my discovery about quicksight’s inability to analyze visuals i was having tons of issues. - the whole ‘calculated fields’ element is terrible and requires you to use their own excel-like syntax for formulas, which would be fine if it weren’t so limited. whereas excel has, what, thousands of functions, and even google sheets has hundreds if not thousands, qs has maybe...100? very constraining having to chain three or four functions to do the job of a basic one in other similar tools. - oh, and to the above, why not allow me to use sql to create a calculated field?!?!!?!?!?!?! this would be SO much easier. - contributing to the overall claustrophobia, there are only like 30 visual styles. that’s more than enough for me, but i have had to compromise some ideas to fit with what’s available. - as i hinted at in my original post, zero analytical capabilities. it can plot fine, but not give you anything to help interpret those plots. utterly insane. - bucketing isn’t possible for a lot of visuals. you have to create a calculated field to say a row falls in a certain bucket, then plot that calculated field vs whatever. very tedious. - it just looks rickety and barebones and clumsy. it’s half a product, at best. e: like, our company's huge project now is going from an infrastructure of snowflake/rds/superset/airflow to an all-inclusive aws package. we've got a contractor working on all of that. and qs is supposed to replace superset as the dashboarding tool. but by god, as bad as superset is, it can at least allow me to analyze data. christ. does anyone know of any free bi tools i might be able to employ given this move? we're only going to get more analytical and i'll be damned if i'm going to calculate INDIVIDUAL loving COORDINATES FOR EACH INDIVIDUAL REGRESSION I WANT IN SQL in order to plot a loving regression on top of a scatter plot. that doesn't just tiptoe over my line, that loving trampolines over it. abelwingnut fucked around with this message at 16:03 on Jan 12, 2023 |
|
#
¿
Jan 12, 2023 15:18
|
|
|
i work with it daily. not sure where you're at with it, but it's basically postgres scaled. i don't know any consultants for it and i constantly struggle to find specifics. to be honest, that's kind of true of all things aws. support and community are very lacking, especially compared to sql server or mysql. also, are we talking redshift serverless or straight-up redshift? biiiiiiig difference in my experience. specific questions about it? i can tell you it's not as good as sql server!
|
|
#
¿
Feb 17, 2024 05:11
|
|
|
|
| # ¿ May 12, 2024 01:17 |
|
|
i'm not a redshift admin, but was previously a sql server admin. i now do more architecture-related things. in order to be the best true redshift admin you can be, you need to know much more beyond redshift. you really need to know like the top 10 aws products, through and through, and know how they interact with redshift. in my experience, the left arm of aws products doesn't talk to the right arm of aws products a lot of the time, and you can easily get caught in situations where your obvious isn't so obvious to others. that said, aws is so vast you can always finagle a solution to whatever problem you might have. but this also demonstrates a fundamental problem with aws. one product doesn't really have the same direction as any other product, so there's almost always some conflict. an example--redshift and quicksight. quicksight is aws's tableau. i feel i should be able query redshift seamlessly. yet, i've run into constant issues doing so. like, you want to run a SET from quicksight? good luck! but that's me approaching it from a sql server logic. if you only know aws and that world, you're probably fine. i don't know. i'm not the biggest fan of aws and would rather stick in sql server. but, and getting to what you're alluding to, aws is 'the future' per the bigger tech companies, no matter how bad it is now. aws is going nowhere, and it's constantly expanding in both smart and stupid ways, so yea, i would say there's a market for database skills in aws. but the job market for tech is kind of lacking already so...might not be the best time to move. abelwingnut fucked around with this message at 06:11 on Feb 17, 2024 |
|
#
¿
Feb 17, 2024 05:54
|
|