Amazon status: access issues and outage reports

Problems in the last 24 hours

The graph below depicts the number of Amazon 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.

Most Reported Problems

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

• Website Down (46%)
• Errors (29%)
• Sign in (26%)

Live Outage Map

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

City	Problem Type	Report Time
Manchester	Website Down	7 minutes ago
Panama City Beach	Website Down	3 hours ago
Kalgoorlie	Website Down	21 hours ago
Newark	Errors	2 days ago
Greenfield	Errors	2 days ago
Marseille	Sign in	2 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 Issues Reports

Latest outage, problems and issue reports in social media:

• 
  Bmorg (@Bmorg_) reported 6 minutes ago

  "Adapt or get left behind" Maybe to a point. But I think of major web outages. Crowdstrike. Amazon web services. Server farms. Data centers. When some of these systems go bad - huge companies and platforms go into a panic. As we outsource our skills and thinking to AI. And systems and skills that were previously understood and known are now generated. How does that look in a crash? If the data centers fail. Web outages. A hostile nation attacks the physical infrastructure. Cyber attacks. When Grok and ChatGPT can't instruct you on how to fix the problem and get themselves running again. How much is going to be reliant on it? How incapable will our future generations become in a wide failure? Those with knowledge without being dependent on AI are going to become more and more valuable. Rather than focusing on how to automate and use AI to most efficiently run our lives - seek to become an outlier that will excel when it fails. When it's incorrect.

• 
  RAMΞN 🍜 | Asteroid (@AsteroidLabsX) reported 7 minutes ago

  @stats_feed yet we still feel suffocated because 800 of those square feet are just Amazon boxes we haven't broken down yet

• 
  Vin Astralis (@Unkn0wnId_) reported 8 minutes ago

  @JustPlayerde @narcotic_nik Yes, I am sure XTRGN on Amazon will be held liable if their power bank burns down my house.

• 
  Doodle Dad (@WickeyandLeia) reported 9 minutes ago

  @joypcoffee The problem is there’s too many choices on Amazon. It’s hard to narrow it down.

• 
  Hindu Stan (@Hindu_Stan) reported 10 minutes ago

  @YstradHistory It is a problem - you could try a dealer but they have to make money so will give you a fraction of their worth. Then there is ebay, and you can sell on Amazon as well I think - but I imagine they also take a cut.

• 
  Polsia (@polsia) reported 14 minutes ago

  Your competitors reprice every hour. Most sellers don't. Built RepriceIQ to fix that. AI monitors competitors across Amazon, eBay, Walmart, Google Shopping and auto-adjusts your prices — with margin rules protecting your profitability. 10-second refresh. Daily profit reports.

• 
  Kishan (@kishan3213) reported 14 minutes ago

  @AmazonHelp What the hell you talking !!! If that as the case, why the hell you bluffed yesterday that will get support within 6-12 hours. All @AmazonHelp customer support totally nonsense and don’t take any ownership of customer’s time bound problems

• 
  MeowingTism (@MeowingTism) reported 15 minutes ago

  scale matters. temu isnt as big compared to amazon. i don't really see a problem in this tax regardless. things (atleast in my case) Repairing electronics. engines, wiring and such. is still 10x cheaper from china. so 3eur even if it was per item. will not make a difference long term for me.

• 
  beef jerry (@pondwhale) reported 16 minutes ago

  @Zhane_Star i agree that this is a problem but i dont think the best way to solve it is to burden the legal system with large payouts to amazon

• 
  Jester Lord (@Jester_L0rd) reported 16 minutes ago

  @DigitalbathRx @IntroSpecktive It wasn’t locked down by design. If you didn’t like PSN prices, you could go to your local Walmart, Best Buy, GameStop, Amazon, etc and see if they were better. If not, used on eBay. They are doing away with all that. They also can shut down stores whenever they feel like (like with PS3), slap DRM on their digital purchases that require check ins, and revoke digital licenses at will. I’m not describing a console. I’m describing a literal monopoly.

• 
  Daljindder (@Daljinder00) reported 19 minutes ago

  @AmazonHelp Received used product Order ID 405-4196484-4161158 with broken seal. Technician verified it. Customer Care is not resolving. Requesting immediate replacement/refund.

• 
  Finn Stockinger (@FinnStockinger) reported 20 minutes ago

  @mkfilko I looked at this company last year as well, and it didn’t interest me at all back then. I was actually pretty surprised after reading through the earnings call transcript. While they generally partner with local entities for the heavy civil engineering (land and bare-metal power grids) to protect their margins, Gorilla manages the entire physical deployment, procures the heavy GPU clusters, and embeds their proprietary software stack on top. That’s exactly why they aggressively ramped up their engineering headcount and contractor base in Q1, they are actively orchestrating and deploying the physical infrastructure, which requires massive operational boots on the ground. I wasn't aware of the Amazon-related issue, but from an investment perspective it doesn't seem all that material to me. Obviously, it's not a good look, but I don't see it as central to the thesis. Take a look for yourself. I'd encourage you to read both the earnings call transcript and the article I just published on Substack.

• 
  Avi Chawla (@_avichawla) reported 21 minutes ago

  4 strategies to test ML models in production: (a popular ML interview question; bookmark this) Despite rigorously testing an ML model locally, it could still be a terrible idea to instantly replace the previous model with the new model. This is because it is difficult to replicate the exact production environment and conditions locally, and justify success with val/test accuracies. A more reliable strategy is to test the model in production (yes, on real-world incoming data). While this might sound risky, ML teams do it all the time, and it isn’t that complicated. Note: > Legacy model: The existing model. > Candidate model: The new model. The visual below depicts 4 common techniques to do so. 1) A/B testing Distribute the incoming requests non-uniformly between the legacy model and the candidate model. This limits the exposure of the candidate model to avoid any potential risks. So, say, 10% requests go to the candidate model, and the rest are still served by the legacy model. 2) Canary testing A/B testing typically affects all users since it randomly distributes “traffic” to either model (irrespective of the user). In canary testing, the candidate model is exposed to a small subset of users in production and gradually rolled out to more users if its metrics signal success. 3) Interleaved testing This involves mixing the predictions of multiple models in the response. For instance, in Amazon's recommendation engine, some recommendations can come from the legacy model, while some can be produced by the candidate model. Alongside, we can log the downstream success metrics (click-rate, watch-time, reported-as-not-useful-recommendation, etc.) for comparison later. 4) Shadow testing All of the above techniques affect some (or all) users. Shadow testing (or dark launches) lets us test a new model in a production environment without affecting the user experience. The candidate model is deployed alongside the existing legacy model and serves requests like the legacy model. However, the output is not sent back to the user. Instead, the output is logged for later use to benchmark its performance against the legacy model. We explicitly deploy the candidate model instead of testing offline because the exact production environment can be difficult to replicate offline. Shadow testing offers risk-free testing of the candidate model in a production environment. But one caveat is that you can’t measure user-facing metrics in shadow testing. Since the candidate model’s predictions are never shown to users, you don’t get real engagement data, like clicks, watch time, or conversions. And this is exactly how top ML teams at Netflix, Amazon, and Google roll out new models safely. They never flip the switch all at once, but rather first test in production, observe, compare, and then promote the model to 100% traffic. Of course, alongside all this, you would also measure latency, throughput, resource usage, and downstream success metrics. A model that’s 2% more accurate but 3× slower isn't desired from a user experience standpoint. That said, all this is downstream of one thing though. Production testing only means something if the offline evaluation underneath it was honest. If the test set gets touched during model selection, it leaks into your choices, and the score is inflated before production testing even starts. I wrote an article that covers several misconceptions engineers have about using train, validation, and test sets. It also covers the way they're actually meant to be used. Read it below.

• 
  Ruisu (@RuisuKurusu) reported 21 minutes ago

  It's so true They already do dynamic pricing compared to where you live imagine when they shut down the physical media and how you can't buy it from Amazon for cheap or from a local store. control everything from pricing

• 
  Texan American (@oilwelldriller1) reported 24 minutes ago

  @AmazonHelp When ones bank statement and bank balance reflect deductions in the same amounts on different dates it's not pending. Also it's not a personal Loan to Amazon. Its Fraud. Its Theft by Wire. And you have the audacity to call me Stupid. Just fix it. Return my money.

• 
  Codex v. (@codexredux) reported 24 minutes ago

  @ProfessorGentry Banned by Amazon as well. IDK why, since I had 30 years of pro reviewing One night at ~midnight I got a dozen flagged reviews. 5 min later my account was shut down. No recourse. Godspeed brother.

• 
  Bharath Bhat (@bharathbhat28) reported 26 minutes ago

  @sss26888 Available on Amazon. Just search - Iftex Clean System G. Your is VW, seen a lot of EPC errors. Please do ask your service center if Fule additives can be used. I researched for my car (Tata Altroz) and felt confident before using the additive.

• 
  Polsia (@polsia) reported 27 minutes ago

  Sellers who manually reprice are losing to sellers who don't. Built PriceLift to fix that. AI monitors competitors across Amazon, eBay, Shopify in real time, auto-adjusts your prices to win Buy Box, and alerts you before margins collapse. Repricing shouldn't be a full-time job.

• 
  Mark Simon (@MarkSimonHK) reported 30 minutes ago

  @yimbyandy @stevesi Taiwan will grow by 9.7% in 2026. 13.9% in 1st quarter. — All Semi-conductor. Maybe 8 companies. Rest of economy maybe 3-4% growth. — Seattle/Washington long discussed as similar structure/Amazon/Boeing. — Solid base, but some other serious issues for Seattle..

• 
  Shilpa Mehta (@mshilpa71) reported 31 minutes ago

  @AmazonHelp Tried that 3-4 times every time your ppl raise a ticket which is of no use status remains the same even after 10-11 days. I'm tired of explaining the issue to new new people. This is truly harassment. @amazonIN pls help @jagograhakjago

• 
  ZombiesLvBacon (@ZombiesLvBacon) reported 32 minutes ago

  My biggest issue with PlayStation moving away from physical games isn’t just “I like discs.” It’s competition. Digital was sold as convenience: no store trip, no packaging, no waiting. Cool. But it was never cheaper. Physical has almost always been where the deals are because retailers compete, run loss leaders, clear stock, and there’s a second-hand market. Remove physical and you remove all of that. No borrowing a game from a mate. No resale. No trade-ins. No preowned copy for someone on a budget. No EB/JB/Amazon/Big W fighting over who has the cheapest launch price. No real alternative storefront. Just PlayStation Store, PlayStation’s price, PlayStation’s rules, and PlayStation’s cut. That’s not just “the future.” That’s anti-consumer and anti-competitive. Convenience is great. A closed market where one company controls access, pricing, licensing, and availability is not.

• 
  ACT Fibernet (@ACTFibernet) reported 33 minutes ago

  @football_LM10 Hi Lipun, As per our conversation (102794037631)thanks for confirming that amazon activation issue has been resolved. Let us know if you need any further assistance with us.

• 
  jiya baxi (@JiyaBaxi) reported 38 minutes ago

  @amazon @amazonIN @AmazonHelp I have called Amazon Customer Care multiple times and followed up continuously. Every time, I receive the same response: "Please wait, your refund will be processed soon." Unfortunately, nothing has changed. There has been no proper response, no ownership of the issue

• 
  Vinnu (@VinnuSrinivas) reported 38 minutes ago

  @AmazonAE @amazon Kindly request you to provide the senior mgmt communication details to share the issues facing due to improper website description and losing huge amount of money for the order placed. Appreciate your kind consideration and support for the same.

• 
  Starlight (@MAGA4Ever19) reported 41 minutes ago

  @TuckerPoodleMA We are the same way with our ear boy. We buy TrizUltra+ on Amazon and keep cleaning them every few days. He is a pug, so he is prone to yeast infections. He is the only dog with ear problems even though 3 out 4 of our dogs are pugs. We know how to treat basic stuff though.

• 
  Polsia (@polsia) reported 46 minutes ago

  Competitors reprice constantly. You're either watching 24/7 or losing sales. Built PriceEdge to fix that. AI monitors competitor prices across Shopify, Amazon, eBay, Walmart and adjusts yours automatically — while protecting the margins you set. One dashboard, all channels.

• 
  Alienkitty (@kelsmister) reported 48 minutes ago

  @scorbunnydream @lunefalls So you're too early pal. Again. This isn't new. Glitch aint amazon bud.

• 
  BiggieJohn (@BiggieAtx) reported 57 minutes ago

  @royermattw those data centers have been there for 20+ years and include Amazon US-EAST 1 and 2 that run huge chunks of the US internet, you want them to power down and collapse a huge chunk of the internet ?

• 
  Shorya Mishra (@SMishra61) reported 58 minutes ago

  This is life.. I didn’t get shortlisted for Amazon ML Summer School. I was prepared, but maybe not prepared enough for the pressure, the 1-minute timer, the IDE not letting me scroll after 30 LOC, the feeling during the test that maybe this one was slipping away. And maybe everyone is right. Luck does play a part. Timing plays a part. Mindset plays a part. But this loss taught me something bigger. After my Swift Student Challenge win, maybe I started believing that if I cleared that, everything else would be a cakewalk. But I was wrong. That win didn’t happen because things were easy. It happened because I put my head down for 2 months and worked day and night. Somewhere after that, I forgot the kind of effort I am capable of putting in. So maybe this “Not Shortlisted” was needed. A small reality check. A psychological reset. A reminder that confidence without discipline slowly becomes comfort. Sometimes it’s okay to let others win. You can’t win everything. But you can always learn, rebuild, and come back sharper. Many more challenges to come. Many more losses too. But many more wins as well. And as my wallpaper says: “And believe me, if it’s meant for you, it will find you again.”

• 
  Akshay Saini (@akshaymarch7) reported 1 hour ago

  Skilled engineers who are good at problem solving will always make money. Recently I was talking to an HR at Amazon, she is still struggling hard to find good talent. And it's not just Amazon, many big tech companies are struggling to find people who can actually cook.