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 (28%)
• Sign in (26%)

Live Outage Map

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

City	Problem Type	Report Time
Paris	Sign in	16 minutes ago
Cognac	Sign in	17 minutes ago
Chhindwāra	Website Down	1 hour ago
Pittsburgh	Website Down	2 hours ago
Manchester	Website Down	5 hours ago
Panama City Beach	Website Down	8 hours 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:

• 
  Ramkishan choudhary (@RKKataniya) reported 5 minutes ago

  Amazon order I ordered American eagle men casual pant from Amazon but I have a size issue so I want to return my order . But I am unable to create return so Please create return Order number 40560177340182704 @AmazonHelp @amazonIN

• 
  7 (@Alpha_Space007) reported 7 minutes ago

  If you order Amazon and cancel an order before it ships they try to steal your money. Its happened to me twice. Both PayPal and Amazon delete the orders. I take snapshots of everything so when they try to do that it doesn't work. Amazon should be shut down. This **** is a scam site. I cant order any large items either especially if its from China they just steal it at customs. Shut the whole ******* thing down Im done fighting for me to get my money back. Thieves!

• 
  Hesper AI (@starhesper) reported 7 minutes ago

  Fable 5 launched June 9. Suspended June 12. Three days. Amazon researchers found the bypass. The new classifier catches it in over 99% of cases. The flagged request is not refused. It is routed to Claude Opus 4.8. No error. A different model. #AI

• 
  Daniel Friedman (@DanFriedman81) reported 8 minutes ago

  @KnightLightXL PSN has a history of movie licensing issues, but Google, Apple and Amazon are much larger platforms for selling digital movies and seem to have them less often, although I am American and these issues seem to affect foreign users more frequently. In order to maintain your T2 Blu Ray, you must keep a Blu Ray player, which Sony manufactures. And at some point, as your Blu Ray gets older and more outdated, they will try to sell you a newer, better disk. Maintaining a collection of aging digital media is not a form of freedom.

• 
  Abhishek Kumar (@AbhiSingh5183) reported 13 minutes ago

  Followed up 25 & 26 June — same script, same prerefund promise, still nothing. Called today, told no ticket was even raised. Asked for written confirmation — flatly refused. Weeks of false assurances, zero resolution. @AmazonHelp fix this. #AmazonFail #ConsumerRights #AmazonIndia

• 
  SuperStraight001 (@Superstraight09) reported 15 minutes ago

  @Retroid0fficial @Zamma3 $70 was pushing it, but $85 after shipping really kills it. I dont know how but shipping has to be improved, maybe make an American store? Went through the same issue with the grips until I found an Amazon listing. I paid $5 more than sticker price but saved tons in shipping

• 
  Kishan (@kishan3213) reported 16 minutes ago

  @AmazonHelp I have already addressed my problem statement yesterday, refer original tweet here. Utterly nonsense

• 
  Stephen (@_88_mph) reported 19 minutes ago

  @AmazonHelp Additionally I have also contacted customer service via the app TWICE and told it was a "technical issue" that has now been fixed each time yet it still continues.

• 
  Malay Krishna (@Malay4Product) reported 19 minutes ago

  A card that handed you 16.5 percent back on Amazon, Swiggy and grocery spends was never going to last forever. On July 1, HDFC finally shut that door. Here is everything that happened during the recent HDFC Smartbuy devaluation. Until now, an Infinia holder could earn up to 15,000 bonus reward points a month by routing everyday spends through SmartBuy gift vouchers. Buy an Amazon or Flipkart or Blinkit voucher on SmartBuy, pay with Infinia, collect 5X points, then spend the voucher like normal. From July 1, HDFC put a separate cap of just 3,000 bonus points a month on brand vouchers bought through GyFTR and Woohoo. The overall 15,000 cap stays, but vouchers can now fill only 3,000 of it. For Infinia that is an 80 percent cut on the voucher route. You now hit the voucher cap at around ₹22,500 of spend, against more than ₹1 lakh earlier. Flights and hotels booked on SmartBuy are not touched, and base points continue. This was coming, and it is the latest measure in a full year of tightening spree, HDFC has been on. Early this year HDFC tried to cut voucher points from 5X to 3X, then rolled it back after the outrage. Then came a redemption cap. Then insurance rewards capped at 10,000 points a month. Then talk of needing ₹18 lakh annual spend or ₹1 crore relationship value just to keep the card. July 1 is the next clamp. HDFC has published this voucher cap for July only so far, so it could get softened like the 3X move did. But the direction across the whole industry points one way. The golden age of Indian credit card rewards is ending. These rewards were funded by the fee merchants pay on every card swipe. UPI, which pays banks almost nothing, ate a huge slice of that volume. At the same time, a small group of sharp users turned Infinia into a 16 percent cashback machine on ordinary spends. The bank was paying out far more than it earned on them. No bank can run that at scale. So every issuer, Axis, SBI, ICICI, not just HDFC, is fencing off the same tricks one by one. Rent. Wallet loads. Insurance. Fuel. Now vouchers. The days of one card printing money for everyone are basically over. From here the smart move is to just pick the card that fits how you really spend, and stop chasing hacks the banks will keep shutting down.

• 
  jiya baxi (@JiyaBaxi) reported 21 minutes ago

  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

• 
  Marlie (formerly yotamy)🏳️‍⚧️🔻 (@DarkMarlie) reported 23 minutes ago

  Had to unfollow this girl cos even though she was nice and fun y she constantly posted the worst selfies I've ever seen. Just absolute dog ****. Awful framing, bad pose, terrible facial expressions, Amazon basics fit, "dreamy" exposure that just comes off as a broken camera. God.

• 
  Dino Bro Don (@DinoBroDon) reported 26 minutes ago

  @SpawnPoiint I’m not arsed, really. I think the only issue I have is the stranglehold they will have on pricing. If you compare the PS store with Amazon for example, the prices of the older games are far lower on disc.

• 
  Energy (@energy_0x) reported 28 minutes ago

  I don't think it has to work that way. People first knew them for food because that's the problem they solved first. That doesn't mean the brand has to stay there forever. Strong brands evolve. Nobody questions when Amazon sells cloud services, or when Uber expanded beyond rides. The product category changed, but the trust people had in the brand made the expansion possible. If Chowdeck can deliver groceries, skincare, and medications with the same speed and reliability people already trust them for, then the brand is evolving from "food delivery" to "on-demand delivery." That's a positioning shift, not necessarily a branding problem. The real question isn't whether they started with food. It's whether they can consistently deliver on the new promise they're making. If they can, customers will adapt.

• 
  Fortyfive Cent 🐐 (@fortyfive_cent) reported 29 minutes ago

  @TheGridPost @Enuguparrot_ what's the issue with the new transformer at amazon street enugu for 2 months now? Do the needful please, is really affecting me as a tailor on that street.

• 
  Tips Excel (@gudanglifehack) reported 31 minutes ago

  Anthropic had to disable these models after Amazon researchers flagged a hacking risk. To fix this, they trained a new protective classifier that blocks the exploit over 99% of the time, redirecting blocked requests to Opus 4.8.

• 
  kiki ✮⋆˙ (@sehaniine) reported 31 minutes ago

  @seerineeee the problem tho is that netflix, amazon, disney and im sure others restrict it to households so i can no longer use my parents streaming services

• 
  🌹 Gween ♥ ACTUAL Fox Yokai 🌹 (@gweeen_tea) reported 32 minutes ago

  @UwUSweetSakura FUN FACT: STREAMING CAN LOOK GOOD! Streaming a 4k Blu-ray rip from a private server and service like Plex or Jellyfin is going to look significantly better than Netflix, Amazon, etc.

• 
  Nunya Business (@CreeperHater42) reported 33 minutes ago

  @DramaAlert I understand his fears, but the way he threw down his hat and put his fists up meant he was having a bad day and looking for a fight. I would have beat his ******* *** and dragged him back to his Amazon truck.

• 
  Rodrigo (@santorfo) reported 34 minutes ago

  @ngxson This is why I hate the fee. The real problem is ******* shein and temu crap but there's so many parts and components that we just cannot get at a good price if we don't own a shop and this is a killer for that stuff. They're just pushing everyone into Amazon at this rate.

• 
  NIGGLER JONES (@144_dooblay) reported 36 minutes ago

  @amazon the fact yall turn the ac off overnight in yall buildings during heat waves is a ******* problem

• 
  Hamid Bergamo (@HamidBergamo) reported 36 minutes ago

  @embodiedthinkr Are you 100% sure about NAC? I heard also some negative experiences such as anhedonia etc - I think it might have some dark side to it. With that said, there was some error on Amazon and I got 10 bottles for 20€ total 😬😬

• 
  Ashley Wright (@ashley_wright) reported 37 minutes ago

  The hardest part of building Social Tale was watching brands quit right before things started working, and it had nothing to do with the work itself. Two to three years ago, TikTok Shop wasn't a priority for most brands. It was more of an experiment. A line item buried under the Meta budget that actually moved the needle. So when results didn't come in the first 30 days, the conversation always went the same way: "We're going to pause TikTok Shop and double down on what's working." I get it, because Meta gives you a clear number. Spend a dollar, get three back, with clean attribution and a short feedback loop. TikTok Shop operates on a completely different timeline. The platform takes time to learn what converts. Creators need two to three posts before they get comfortable with a product. Algorithm momentum builds over weeks, not days. Trying to educate a brand on that while Meta is sitting there with clean dashboards and predictable returns is one of the hardest sells in this industry. It's like telling someone to get on Amazon in 2008. The opportunity is obvious in hindsight. In the moment, it just feels risky. The brands that stuck with it are now compounding. More creator relationships, more algorithm momentum, more GMV. The ones that paused are starting from scratch against competitors who never stopped.

• 
  Avi Chawla (@_avichawla) reported 38 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.

• 
  BXG (@Lawgiver771) reported 40 minutes ago

  @briananderson58 @_ToastedBrioche @YaboiHoxen Digital isn’t the problem; PlayStation is. Bad refund policy, multiple security breaches, and first-party games stay overpriced on the PS Store way longer than physical copies. Compare Ragnarok on Amazon or Walmart to the PS Store.

• 
  Star Wars Timeline (Ben) 🇷🇺🇺🇸 (@SWT_Channel) reported 41 minutes ago

  @The_Top_Loader @VGMHero Same here. Locked down my Amazon preorder with 6 games!

• 
  🌹 Gween ♥ ACTUAL Fox Yokai 🌹 (@gweeen_tea) reported 42 minutes ago

  @UwUSweetSakura AND STREAMING CAN LOOK GOOD! Streaming a 4k Blu-ray rip from a private server and service like Plex or Jellyfin is going to look significantly better than Netflix, Amazon, etc.

• 
  ACT Fibernet (@ACTFibernet) reported 42 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.

• 
  PlayOJO (@PlayOJO) reported 47 minutes ago

  We left a coin with these Bucks and we can't find it! 🫢 ⁣ ⁣ Help us track it down and you could 𝐰𝐢𝐧 𝐚 £𝟓𝟎 𝐀𝐦𝐚𝐳𝐨𝐧 𝐯𝐨𝐮𝐜𝐡𝐞𝐫.⁣ ___________ ⁣ Buildin’ Bucks by Play N Go. Three randomly selected winners will be announced on 21st July to win a £50 Amazon voucher each. Vouchers are not exchangeable.

• 
  GUL (@gulVasikova) reported 56 minutes ago

  $LUNR Intuitive Machines (LUNR): This Is Bigger Than One NASA Contract Most investors see a $148 million NASA contract. Long-term investors see something much more valuable. They see another brick being added to America’s lunar infrastructure. Think about Amazon in the early 2000s. People focused on every online order. Very few realized Amazon was quietly building warehouses, fulfillment centers and logistics networks that competitors would spend decades trying to replicate. That’s what Intuitive Machines is doing today. The Nova-C lander isn’t just delivering scientific equipment to the Moon. Every successful mission improves landing precision… navigation… communications… surface operations… and gives NASA more confidence to award future missions. This latest contract also shows something important. NASA isn’t simply buying a rocket launch. It’s building a long-term commercial ecosystem where private companies become permanent partners instead of one-time contractors. Every successful mission strengthens Intuitive Machines’ reputation as a trusted lunar logistics provider. That trust is extremely difficult to buy. Yes, investors should acknowledge the risks. The company is still unprofitable. Insiders have been selling shares. Execution risk remains high because space missions leave very little room for error. But many early infrastructure companies looked expensive and unprofitable before their networks became valuable. The real question isn’t whether this single contract adds $148 million. The real question is whether Intuitive Machines becomes the “FedEx of the Moon.” If NASA continues returning to the Moon throughout the 2030s, every mission completed today could become a reference that wins multiple contracts tomorrow. Sometimes the biggest asset isn’t today’s revenue. It’s becoming the company governments trust when failure isn’t an option.

• 
  Opeyemi Olutayo (@OpeyemiAxra) reported 1 hour ago

  Read Amazon updated policy on June 30th my friend, all VIRTUALS BANKS has been frozen out to ACCEPT PAYMENT e.g. Royalties but they can do other payments on Amazon. To save you the stress, see the attaches for the policy reviews and the PYHSICAL BANKS you can use. It is not a APP PROBLEM, IT IS A POLICY PROBLEM. Wishing you the best. Keep using @useaxra