Latest posts tagged with #AWSfail on Bluesky
*selects 11,000,000 strings into a sorted text file
it's takin a while
jesus, fuck, crashed the jumphost with this? HOW
fucking EBS again
i have to deal with this SO MUCH
if you hit an EBS quota of iops or throughput it KILLS THE MACHINE
things are SO FUCKING STUPID now
#aws #awsfail #ec2
screen grab from funny parody video wizards discuss system failure
3 out of 4 people in my family will find this excruciatingly funny and poignant.
#enshittification #awsfail #AIisanechomachine #worksmart #plandon'thack
www.instagram.com/reel/DKfuAiw...
"it seems that the talent who understood the deep failure modes is gone. [New] teams lack the institutional knowledge needed to [prevent incidents or] reduce the time to…recovery. …AWS' operational strength was built on redundant, experienced people, and when you cut…things start breaking."
#awsfail
I guess we need to stop calling this stuff "smart"
www.dexerto.com/entertainmen...
#aws #awsfail hat tip @eikonos.bsky.social
Thank you Amazon and AWS for making my class day where, I have no students today, moderately useless up to this point. #AWSFail
Screenshot of chart from AWS ECS service health panel, "Memory utilization", showing Minimum, Average, and Maximum traces in Blue, Orange, and Green, on a charcoal, dark mode background. The Y axis is labelled 10.2 (top), 10.1 (midddle), and 9.96 (bottom). In other words, the memory is fluctuating by < 2% despite the mountainous looking graph. Basically misleading, at a glance.
AWS doing its Y-axis thing (unfixed for many years). #awsfail
Screenshot of htop display on a Linux server, showing about 32 processes dispatched by GNU parallel. Each process is running a single aws s3 rm (which happens to allow spec of only a single object anyway). The machine is pegged at 100% x 8 cores.
I was stunned when I first saw this diagram and it hasn't changed in the months/years since they rolled it out, but then they never fix anything bsky.app/profile/symb... #aws #awsfail #awsconsole #ux
Screenshot of AWS Console showing the VPC "resource map". Left column: Subnets Middle column: Route tables Right column: Network connections There are lines drawn (with curved corners! futuristic) between subnets and route tables, and between route tables and network connections (gateways). However, the lines are all the same grey colour and run over each other, so it's impossible to read how they are actually connected. (Unless you hover.)
Where is the adult supervision at AWS? Why would you _deliberately_ obfuscate the connections when you have about eleventy million pixels you could use to clarify them?
Typical #awsFail #ux
I have 900 messages to download from S3!
This wouldn't be a big drama, Except that no matter where you use it from, `aws s3 cp` takes around 1 second PER OBJECT no matter how small
#awsfail
lol rds now tries upselling add-ons during the long wait for database creation
and has a 'hide for 30 days' checkbox
next: "Pay $3 to hide this modal for 30 days"
#aws #awsfail
Chart from AWS, Lambda Monitoring tab. "Error count and success rate (%)" is the title, and yes, those are the two quantities plotted, somewhere. Y axis is labelled "Count" on the left, with ticks at 0, 0.5(!), and 1; On the right it's labelled "No unit" (but apparently it's the Success rate percentage), with ticks at 99% (!), 99.5, and 100.
It's almost comical how bad the charting remains.
#aws #awsfail
Screenshot from Elastic Beanstalk section of AWS Console, Monitoring tab. There are 2 rows of three panels, each with heading, number, and chart: • Environment health, "15" (…okayyyy) • Target response time, "0s" (seems fast!) • Request count, "1" (the chart kind of contradicts this) • CPU utilization, "0.9%" (ok, when?) • Network In, "202 kB" • Network Out, "9.62 MB"
I defy anyone to find meaning in the very bold and large numbers
Or indeed, the charts, since there is no axis labelling
#awsfail #elasticBeanstalk ← Never, ever use this product
2) HOW DO YOU "RUN OUT" OF VIRTUAL INSTANCES YOU CLUSTERFUCKS.
#awsfail
#darkpatterns
ok so I remade the cluster in the way that gives you a supposedly correct ASG, instance template, etc, …and it still doesn't work lol #awsfail
they do this instead of hiring people to improve documentation. several years ago, a less jaded version of me sent an aws console enhancement idea to support: link each view/form to the relevant doc section for the product. they responded enthusiastically but never did anything with it. #awsfail
Screenshot of AWS Console Secrets Manager page advertising a storage cost of "$0.40 per secret per month and $0.05 per 10,000 API calls"
lol $0.40 per secret per month #awsfail
Screenshot of AWS Console CloudFront monitoring chart showing "Origin Latency". The Y axis has 3 labels, "153", "9.66k", and "19.2k". No units given...
New #awsfail thread
old one at the bad place, twitter.com/Symbo1ics/st...
AWS Console warning message in the Route53 interface, while trying to add an A/Alias record to a CloudFront distribution (it won't let me choose one): "An alias to a CloudFront distribution and another record in the same hosted zone are global and available only in US East (N. Virginia)." (I don't think this warning is related to why I can't select a CF distribution though. This zone already has three aliases to various CF distributions, why can't I add a fourth?)
