1. Home
  2. Companies
  3. Amazon Web Services
Amazon Web Services

Amazon Web Services status: access issues and outage reports

No problems detected

If you are having issues, please submit a report below.

Full Outage Map

Amazon Web Services (AWS) offers a suite of cloud-computing services that make up an on-demand computing platform. They include Amazon Elastic Compute Cloud, also known as "EC2", and Amazon Simple Storage Service, also known as "S3".

Problems in the last 24 hours

The graph below depicts the number of Amazon Web Services reports received over the last 24 hours by time of day. When the number of reports exceeds the baseline, represented by the red line, an outage is determined.

At the moment, we haven't detected any problems at Amazon Web Services. Are you experiencing issues or an outage? Leave a message in the comments section!

Most Reported Problems

The following are the most recent problems reported by Amazon Web Services users through our website.

  • 50% Website Down (50%)
  • 31% Errors (31%)
  • 19% Sign in (19%)

Live Outage Map

The most recent Amazon Web Services outage reports came from the following cities:

CityProblem TypeReport Time
Kyiv Sign in 6 days ago
Chennai Website Down 9 days ago
Point Pleasant Beach Website Down 13 days ago
Little Rock Errors 17 days ago
Atlanta Website Down 17 days ago
Clearwater Website Down 18 days ago
Full Outage Map

Community Discussion

Tips? Frustrations? Share them here. Useful comments include a description of the problem, city and postal code.

Beware of "support numbers" or "recovery" accounts that might be posted below. Make sure to report and downvote those comments. Avoid posting your personal information.

Amazon Web Services Issues Reports

Latest outage, problems and issue reports in social media:

  • yiyine8
    いいね! (@yiyine8) reported

    @AWSSupport Urgent billing issue. My EC2 instances are still generating charges, but I cannot access my AWS accounts because the registered emails were deleted. Case ID: 177725454200231 Could someone please help review my case?

  • mattflashuk
    Matt Adams (@mattflashuk) reported

    @AWSSupport 40 days for a billing issue. Can you look into this please.

  • oraclee_xyz
    oraclee.btc 🧠 (@oraclee_xyz) reported

    @awscloud @OpenAI but can it fix ****

  • Witherndale99
    Nigel (@Witherndale99) reported from Oakville, Ontario

    @awscloud Human interface really sucks right now. Its difficult to use, error prone, lacks 99% of the users knowledge and context. So, a non invasive, adaptive interface, capturing input without contact, and also the ability to access my knowledge, experience, thoughts, and emotions

  • RobinDavidMorri
    Robin Morris (@RobinDavidMorri) reported

    @AWSSupport @UttamTarasariya think there's a general outage bc I can't login either. passkey not working and SMS message not delivered when I try MFA.

  • St3v3Wiederrich
    Steven Moerder (@St3v3Wiederrich) reported

    Did someone hack/hijack the Waymo server? @awscloud does waymo use anything on cloud from you? @GoogleDeepMind or other services? @Waymo what is this? This looks like some dystopian nightmare fuel. Where someone pretending to be a computer or hacker did this maliciously?

  • Chrispractical
    practitioner (@Chrispractical) reported

    @awscloud The real advantage isn’t budget size… it’s how fast you can safely ship AI that actually solves real problems in regulated industries

  • RaviYadav514011
    Ravi Yadav (@RaviYadav514011) reported

    I returned a product from Amazon online shopping app and the product was not working so I returned it but My refund is not released yetsince 2 months worst app @amazon @amazonIN @awscloud

  • pelaseyed
    homanp (@pelaseyed) reported

    @0xMevu @specific_dev @awscloud It's a swarm of researches, we don't have issues with tool calls.

  • imTkorde
    Tejas Korde (@imTkorde) reported

    @AWSSupport hello AWS despite asking for multiple times to connect with the customer care executive I'm just getting we are working on it and I have to solve the issue today only please connect with the team it's highest priority task !

  • vjswami
    Vijay Swami (@vjswami) reported

    In an Agentic AI sales cycle, you need to intimately understand current customer workflows before you can even think about agents. This concept of deeply understanding the problem before "solution'ing" was something drilled into our heads at @awscloud in the early days. Prior to the current tools, this involved lots of back/forth with a customer, then having to put together a working document in the shape of the workflow for that particular customer and/or department. Time consuming and error prone. Today? Another day, another @AnthropicAI @claudeai Claude Code skill developed to speed this along: /workflowdiscovery It can take in multiple types of unstructured data (emails, docs, images, etc) and then turn them into a nice living document that maps the workflows, stakeholders, inputs/outputs, artifacts and more... in minutes. This is then automatically hosted on an internal website and the sales team gets a link they can use to go back to the customer and live edit the doc Q&A, validate assumptions, etc to lock this down. And then this can feed right into the solutions/engineering team with a simple JSON export into their workflows for agent mapping. Below is the skill information on the left, and sample output on the right. These are the types of tools and workflows which are mandatory for sales teams in 2026+.

  • Nma_ci
    Marci (@Nma_ci) reported

    @AWSSupport @AWSSupport @awscloud We are currently having an issue on our Production on AWS because a required quota is not being approved. This same quota was approved in our Test Environment on AWS. Support Ticket keeps getting closed. caseID: 177910350200072

  • PhilD4064929573
    PhilD (@PhilD4064929573) reported

    @RealAllinCrypto @coinbase @awscloud Coinbase shld wake up & partner with $ICP & they wldnt be down right now costing them millions.

  • Abomination81
    Abomination (@Abomination81) reported

    @paulonoldin Ahh there is a huge problem. get an amazon aws in ireland, its like 5-10 a month usd. It will massively improve it

  • JodhaTarun60774
    Tarun Singh (@JodhaTarun60774) reported

    @AWSSupport Still not working...

  • maynor_8795
    Maynor Henríquez (@maynor_8795) reported

    @AWSSupport @AWSSupport Still NO resolution on Case 177754503800193. I’ve lost thousands of dollars in halted operations, delayed international shipping, and offline medical platforms. The lack of urgency from your Trust & Safety team is absolutely shameful. Fix this immediately.

  • TeriRadichel
    Teri Radichel #cybersecurity #pentesting (@TeriRadichel) reported

    I’m using @anthropic opus 4.8. With @awscloud Kiro. I have rules in my README that say only use Bash. But it ignored my instructions and said sorry I used Python after using a lot of tokens to do something I didn’t want it to do. I feel like I should get those tokens refunded: this is the kind of thing that unexpectedly wastes money and time and tokens unnecessarily. I can’t make the instructions any clearer. Stuff like that is so annoying. But I’m still writing code way faster. The problem is I got a $20 plan and used it up in a few hours. I tried switching back to 4.6 but seemed like the same results for code written. It’s like everything got more expensive after the new model showed up. Or maybe I really was just making that big of a change. I had to just up the subscription to $40. I’ve already spent a bunch of money on other subscriptions and got incorrectly billed $200 in tokens I couldn’t use this month (working on getting that refunded). So you can do the math. $40 per day x 31 = $1240. But people are using multiple agents. I could spin up an army of agents to write code for me if it didn’t cost so much. Who is going to win this game? As a single developer I’m seeking ways to reduce token usage (or whatever the billing metric is) and mistakes by agents like this one feel so unfair. It clearly was not what I asked for. Whether I fix it or the agent notices its mistake after the fact and fixes it without telling me it’s still going to cost the same. Bottom line is that I can produce a lot but it also costs me a lot and so far I haven’t found a way around that besides writing deterministic code. I have to get the thing built I want to make me money before I spend it all on AI tokens. But right now the agent is making mistakes and getting stuck on things I’m doing that I guess are unique and the agent hasn’t seen too much before. It writes code and can’t figure out what it did wrong, AWS shared networking with VPC endpoints and organizations delegated admins and such I’m getting around that (without giving up full access to my cloud account) by creating verification queries and code but then the agent doesn’t follow the existing file structure and uses the wrong language without at least asking…*sigh* Also it seems like the verification code is burning tokens like a wildfire out of control. Not sure how this is going to work out. I hope there’s not some issue like someone else siphoning off tokens behind the scenes or some other weird problem like the previous nerf issue.

  • WE2SeriousConvo
    TechBrosTechBro (@WE2SeriousConvo) reported

    @JeffBezos We need to give everyone in the world 1VCPU for free for life from amazon aws. That will solve all world problems. People can rent them out or host their websites and or business ideas. Call me Jef..

  • akashtattva
    akash (@akashtattva) reported

    i wonder whether anyone has ever mapped amazon AWS server costs to specific product feature improvements. because this kind of accounting framework being applied ai spend seems intuitively wrong to me.

  • odyhub
    OdysseasPapadmitriou (@odyhub) reported

    @AWSSupport - Redis issues on Virginia US-East1 - why are your updates so incomplete?

  • JustABoyLost
    Saurbh Pandey (@JustABoyLost) reported

    @AWSSupport @AWSSupport The replies are coming too slow and my business is getting impacted coz of this

  • TeriRadichel
    Teri Radichel #cybersecurity #pentesting (@TeriRadichel) reported

    I’m using @anthropic opus 4.8. With @awscloud Kiro. I have rules in my README that say only use Bash. But it ignored my instructions and said sorry I used Python after using a lot of tokens t do something I didn’t want it to do. I feel like I should get those tokens refunded: this is the kind of thing that unexpectedly wastes money and time and tokens unnecessarily. I can’t make the instructions any clearer. Stuff like that is so annoying. But I’m still writing code way faster. The problem is I got a $20 plan and used it up in a few hours. I tried switching back to 4.6 but seemed like the same results for code written. It’s like everything got more expensive after the new model showed up. Or maybe I really was just making that big of a change. I had to just up the subscription to $40. I’ve already spent a bunch of money on other subscriptions and got incorrectly billed $200 in tokens I couldn’t use this month (working on getting that refunded). So you can do the math. $40 per day x 31 = $1240. But people are using multiple agents. I could spin up an army of agents to write code for me if it didn’t cost so much. Who is going to win this game? As a single developer I’m seeking ways to reduce token usage (or whatever the billing metric is) and mistakes by agents like this one feel so unfair. It clearly was not what I asked for. Whether I fix it or the agent notices its mistake after the fact and fixes it without telling me it’s still going to cost the same. Bottom line is that I can produce a lot but it also costs me a lot and so far I haven’t found a way around that besides writing deterministic code. I have to get the thing built I want to make me money before I spend it all on AI tokens. But right now the agent is making mistakes and getting stuck on things I’m doing that I guess are unique and the agent hasn’t seen too much before. It writes code and can’t figure out what it did wrong, AWS shared networking with VPC endpoints and organizations delegated admins and such I’m getting around that (without giving up full access to my cloud account) by creating verification queries and code but then the agent doesn’t follow the existing file structure and uses the wrong language without at least asking…*sigh* Also it seems like the verification code is burning tokens like a wildfire out of control. Not sure how this is going to work out. I hope there’s not some issue like someone else siphoning off tokens behind the scenes or some other weird problem like the previous nerf issue.

  • AstroKanu
    Astro Kanu (@AstroKanu) reported

    @awscloud @AWSSupport it’s a live project impacted - ur disrupting my clients- pls stop this casual approach and take this on priority- I am not able to login to my AWS account so I can’t access ur link - I’ve screen shots of all payments done which I can share resolve this now!

  • pactcg
    Pacific Card Exchange (@pactcg) reported

    @AWSSupport Yes. Have a case up online but cannot get in contact with anyone. Still dealing with my site down for the past 2 days w/o any notification from AWS.

  • ankurshn
    Ankur Shinde (@ankurshn) reported

    I went to the @awscloud in Mumbai yesterday. I'd been once before, and the reason I go is simple: AWS is the largest cloud player, and I want to see where the industry is actually heading and not where the blog posts say it is. I'll be honest about how the day started. Registered attendees were not let in until 1:30 PM. I reached by 11 after a train delay, planning to walk the AWS Village first. Instead I stood outside for two hours in 45°C heat. I wasn't alone - a sponsor next to me had flown in from the US and told the guard he'd paid $50,000 to be there. Didn't matter. The guard was honest and only following orders, so I don't blame him; the problem sat higher up. A few presenters gave up and left for other work, so the schedule on the site stopped matching reality. Once I got in, I was deliberate about where I spent the time. I went straight to the AWS for Financial Services booth - my first time there, and I'd planned for it. I saw the merchant onboarding flow, loan processing that runs in about three seconds once you upload your documents, and how customer-interaction data gets fed back to sharpen their recommendation models. This is the part of finance I actually care about. Then the Anthropic session, which was the one I came for. Three things stuck with me. // India is the second-largest country using Claude Code, behind only the US. Given that we produce most of the world's software engineers, the size of the gap that's still open says more about the opportunity than anything else I heard that day. // Business and finance was barely 2–3% of all usage, far behind computer and mathematical work. I walked in expecting that number to be high. It wasn't, and that surprised me more than any of the headline stats. // And the point I keep coming back to: agents are designed to ask for help on purpose. We say we want full autonomy, but the engineering choice to make an agent stop and request human context is exactly what makes it usable. I hadn't thought in that way before. Three things speaker (Sorry Idk the name :) said to act on: • Pick the boundary problem, • Invest in context and capability, • Start where you are - you don't need to reinvent anything. One last thing, nothing to do with the summit. On the train back to Nashik, a senior from my college got on one station after me - he's interning in Mumbai and was also heading home. Completely by chance, same train, same destination. Small world.

  • TransFemPyro
    FemPyro 🏳️‍⚧️ 🔞 (@TransFemPyro) reported

    FWI Amazon AWS is down so its not really VRChats fault. This also means other services like tiktok are down.

  • HereComesJDOG
    It’s Boofing Time (@HereComesJDOG) reported

    @william_fitz @awscloud is going to hit a wall with this as well. They’re practically forcing AI solutions down customers’ throats, it’s crazy.

  • RealHaraldJezek
    Harald Jezek (@RealHaraldJezek) reported

    @AWSSupport It's really frustrating. No matter how I get in touch with you people, there is no attention to my problem. This issue is now going on for already 5 days and there is ZERO response.

  • zippy257
    Zippy (@zippy257) reported

    @awscloud @OpenAI wake me when one of these agents actually survives a real outage without human rescue

  • oSkyUpo
    عبدالله (@oSkyUpo) reported

    @awscloud Middle East FC 26 Pro Clubs players are being routed to EU servers instead of Bahrain/UAE servers, causing high ping and delay. The issue is widespread across Gulf players. Please investigate.