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 (47%)
• Errors (28%)
• Sign in (25%)

Live Outage Map

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

City	Problem Type	Report Time
Southampton	Errors	8 minutes ago
Valencia	Errors	54 minutes ago
Les Herbiers	Sign in	2 hours ago
Coacalco	Sign in	3 hours ago
Paris	Website Down	4 hours ago
Rouyn-Noranda	Website Down	6 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:

• 
  Tiji George (@george_tiji) reported 4 minutes ago

  @AmazonHelp Well rather than finding issues with customer and giving silly excuses, why can't u focus to investigate the following: 1) why was the ordered cancelled last minute and then price increased 2) why was the cancelled product delivered almost after 10 days

• 
  boris spider (@John_C_Spring) reported 4 minutes ago

  @Daniel_Rolands I loved the book series. Watching the Amazon series right now. Super gay and jeety. What a failure for a story that beat GOT hands down. Jordan is rolling in his grave you woke bunch of jackasses.

• 
  Rudra_Shankara (@blackstrome456) reported 6 minutes ago

  If any one Buyed larger Item anything from @amazonIN and seller sending it with @delhivery try to cancel it that time only they Play with your time, But will Delivery it and make Fake excuses that customer not present and some problem occurred after 7 to 8 days amazon will refund

• 
  intellivision (@intellivisi0x) reported 6 minutes ago

  @thierryntoh23 @25YearsAgoLive Amazon? Their stock is down like 98% in the last 18 months, i don't think they're going to make it. Haven't heard of Netflix, but cool name I guess.

• 
  Pranav (@damaniya_pranav) reported 12 minutes ago

  @amazonIN Posting this tweet to bring this issue to your attention. I hope I'll get back my money from Amazon

• 
  Gaurav kumar (@gaurav0721) reported 16 minutes ago

  @AmazonHelp Please Read the thread properly and also in message all the issue has been explained so please do the needful on priority

• 
  cherie (@imherenow1) reported 19 minutes ago

  Governments are for profits, not people or food security. We’re losing farmland at alarming rates Some land has been expropriated. Some farmers have given up, without supports. This is a national security problem as we can’t eat Amazon warehouses! Food security is at risk

• 
  🔻agitprop + absurdity🔻 (@agtprpnabsrdty) reported 20 minutes ago

  Iranian Shahed drones struck two Amazon Web Services data centers in the United Arab Emirates before dawn on Mar 1, sparking fires and knocking out banking apps for 50 million people, confirming that Gulf AI infrastructure has become a live wartime target. The strikes turned commercial data infrastructure into a target within months of construction. The UAE's G42 became one of only two non-American firms on the Islamic Revolutionary Guard Corps' target list, a status it earned by phasing out Huawei hardware and divesting Chinese holdings as a condition of receiving US chip export licenses. The same AWS infrastructure hosting Gulf banking and civil services was simultaneously processing targeting data for Operation Epic Fury, the US-Israeli campaign against Iran that Gulf governments had neither endorsed nor been consulted on. Gulf states passed data localization laws requiring sensitive state data to stay inside national borders, sold as protection from foreign dependence. When Iran hit the AWS facilities, governments found themselves legally barred from rerouting workloads. Aligning exclusively with the American chip stack meant Chinese cloud providers, the one available diversification route, had already been cut out. Saudi Arabia pledged a trillion dollars in US investment for the right to buy Nvidia Blackwell chips through Humain, while the UAE anchored Stargate UAE, a 5 gigawatt campus spanning 7.3 square miles of Abu Dhabi desert. Washington granted the UAE license-free access to advanced chips for Amazon, Microsoft, Google and OpenAI. None of that capital protected the population living next to the hardware. Nine in ten Gulf residents are migrant workers with no stake in the confrontation, and they spent March unable to pay for taxis or access banking apps while their governments were shut out of both the war planning and the ceasefire talks that followed. My take: Gulf monarchies traded oil sovereignty for compute dependency. Riyadh and Abu Dhabi built the most expensive server farms outside the United States and discovered that hosting them turns their territory into US military infrastructure. The chip licenses behind Stargate UAE and Humain came stapled to a loyalty test measured in divested Huawei gear, and Iran treated that as grounds for war.

• 
  Brian Bakker (@BrianBakker62) reported 20 minutes ago

  The project is still alive, just parked until the dust settles. In the meantime, *Byte 01: A Spark of Logic* is still out there holding down the fort on Amazon and the website. (Link below). For now, it's back to my primary hustle these days.

• 
  Vasu Battula (@UltimateVasu) reported 26 minutes ago

  @AmazonHelp Miss indhu, This issue is still unresolved. Over that everytime i try to contact through mail, your agents keep changing and everytime a different thing is being said.

• 
  Hello You Experiment (@HelloYouExp) reported 27 minutes ago

  projLog: LinkedIn primed, system status? Genotype written, phenotypes WIP. How are the custom site and UI interactions going? I dropped it. Like the Huemanifesto not having a hover state on mobile and it uses a Figma Sites template, I don't want AI to play a huge role in the presentation of the content. But it would be so sick to have translation windows, variable views, animation, and visua... No. A choice was made not to use AI for mass deployment or automate the critical thinking, it isn't consistent to use it for flashy visuals to explain the content. You really are not fun sometimes. Do you want to speak to audience needs, wants, or desires? None and all. I do not want to sort one higher than the other, that would just create viral content. Each person has to see themselves in the story, part of the struggle. Chirality is going to require work from the user though, decreasing adoption probabilities. I'm confused. Making it hard to understand, translate, interpret, read through, and find is going to decrease the total number of conversions, the metrics will be low. Did you not see the TPS report this morning? No, I didn't get mine, where would it be? It's on your desk right there <points to blue binder > What!? That's blue, the TPS report is red, it's always yellow. < puts coffee down; hands start animating > Why the hell would they change the cover to blue? I didn't know it was there. I come in here ten minutes before 8 am, make my coffee and sit down to read my TPS report. I don't have to wonder around the office searching and asking for help because my TPS report is yellow ******* yellow, it is always yellow and it is right here, everyday. They can't go changing the color system, why change something if it works??? < coworker passes by and randomly comments but never stops their forward motion > Slavery worked. Whoa, whoa, Claire, that's not, I'm talking about binder colors, not, that's...<turning back to original conversation>...that has nothing to do with this, I should report her to HR. She is just making a point, you were comfortable with the previous system. Don't read too deep into it. I just don't understand why change to the blue covers, I don't want to think at 8 am, I just want to do my job. That's an interesting way to phrase that considering our company mission statement is "change the world to think differently". < picks back up coffee > Great pivot, let's back into the conversation that's most important. How can we increase the adoption metrics for HYE across the board? I want to end Q3 with a win and take the trophy away from the customer service team. That's your personal goal? A three-foot yellow trophy that they bought on Amazon to facilitate higher internal output? **** those guys, they've won three times in a row for "customer service", like that's hard. < eyes drift towards the ground, passionate obsession > Our team deserves that trophy, we build the product that makes all the money, all these other teams are second at best. < coworker passes by and randomly comments but never stops their forward motion > Uh huh, that's what they said back then too. < sets down coffee > Whoa, whoa, Claire, that's not, we are not even talking about colors, not, that's...<turning back to original conversation>...you know me, that's not me, and I...that's just offensive, I support everyone equally, very inclusive, I have friends that.. <voice fades as Claire disappears down the hall, never looking back > < interrupting > I need to get to work, this HYE project isn't going to connect itself, you know since we chose not to use AI agents. Okay fine, but how do other companies make content highly sharable, like what's their secret? BuzzFeed creates fast, highly shareable content that compresses complex ideas into quizzes, lists, simplified identities, and emotionally optimized headlines. But over time, it reduces tolerance for nuance, rewards reaction over understanding, and trains people to consume news, culture, and identity through the same attention-driven logic. That's great design! They have amazing metrics! Let's do that! < points to customer service > You're going down! No, the name "Buzzfeed" originally described their experimental algorithm, which tracked what topics were creating a "buzz" and fed them aggressively to readers. So we are like, doing the exact opposite of what everyone proved to be highly successful? Do you not want to win? What are we even doing here if you don't ******* want that yellow trophy!? I really should get back to work... I'm going to be honest with you, as a friend, you're what, like 40-something and you're still an individual contributor. This company needs leaders who can win, win, win and if that isn't you then maybe you don't want to change the world. < gen z stare > I'm serious, and as the leader of this team, I need someone I can count on to make those metrics rise up and do their little dance, so I'll ask nicely, whatever color folder they decide to put on my desk for the TPS report, just change it out to a yellow one before I get in and make the design easy so we can all win. Just do the job we gave you and maybe in the future you'll have the freedom to pick... < looks over; Claire heading back towards them and is making eye contact > ...uh, to pick whatever design principles are important to you... < grabs coffee and scuffles back towards cubicle > ...Q3 is about winning! Let's get it done! < quickly ducks into cubicle > < coworker passes by with a purple colored binder and randomly comments but never stops their forward motion > Another day, another dollar, right? At least they pay us < wink > Claire, do you even know what we were talking about? < she pivots to face them, seamlessly transitioning into a backward stroll > With your boss? Everyone who is told to uphold something they don't believe in has the same facial expression. Did you learn that from "Don't Lie To Me" on TV? What does my face say right now? < she turns forward, talking over her shoulder as she walks away > Ha, that you mean business. Heads up, tomorrow's TPS report will be in a purple binder. < a voice explodes from a nearby cubicle > < clears throat; heightened voice > Claire, where are the old yellow... < Claire's voice fading > Anyone clinging to the old system can find the past in the archives room. Hmmm.. < glancing over at the poster on the wall > "Change the world to think differently"

• 
  Igor Tikhomirov (@IgorTikhomiroff) reported 27 minutes ago

  @FinancialPhys That’s not how it works bud. If dollar stores could pay a bit more to keep the employees they would. Everyone is going down my man as Amazon and Walmart killed everything

• 
  PrinceLD Martian (@PrinceLDCharley) reported 32 minutes ago

  @AmazonHelp @JeffBezos @ajassy I need an immediate escalation to someone who can reconcile the written promises from your support team with this final rejection. This is a system error based on conflicting information provided by Amazon.

• 
  Shawn (@raptorspike) reported 35 minutes ago

  @rez_devil @amazon @UPS I've only had one issue in the last two years where something was delayed more than a few hours. Package went missing completely in their system. They reshipped it, the both the new and original arrived within a day. Got to keep both for the cost of one

• 
  Victor (@MotherDragonBSV) reported 36 minutes ago

  Let alone if people realized that real Bitcoin is $BSV precisely because of the big blocks unbound and scaling with no limit to be P2P Electronic Cash + A new system built for micropayments and and internet of Value moving at 1M-1B+ transactions per second. $BTC isn't an email's inbox that has a "spam" problem. It's problems are that: - It's pretending to be the real Bitcoin Protocol $BSV - It's pretending to be "digital gold." - It's pretending to be "savings technology." - It's pretending to be "a hedge." - It's pretending to be "a store of value." - It's pretending to have "utility/use cases." - It's pretending to be "sovereign." - Nodes that don't mine are pretending to "verify" so they feel special too. - It's pretending to be "decentralized." - It's pretending that "cryptocurrencies, L2's, and payment channels have any ability to scale." They don't. - It's pretending that there is no "second best," when it in fact isn't even second best to Gold and the Stock Market. HELL it's currently #14 for what's pretending to be "an Apex Asset" underneath: - Tesla - Meta - Suadi Aramco - Broadcom - SpaceX - TSMC - Amazon - Microsoft - Silver - Alphabet (Google) - Apple - Nvidia - Gold If companies can go out of business, then what make you think the HODL cult Ponzi that is "BTC" pretending to be a whole lot of things... can't fold? Spam isn't BTC's problem. BTC problem is that the very idea of BTC is the SPAM. SPAM = "unsolicited, unwanted, or repetitive DIGITAL COMMUNICATIONS sent in bulk to a LARGE NUMBER OF RECIPIENTS. Typically, used for advertising, phishing scams, or malware distribution, it most commonly appears in emails, text messages (SMS), social media, and forums." BUT CAN SOMETIMES ALSO appear as LASER EYES maxis of the BTC HODL cult selling GREATER FOOLS into buying their BTC bags that desperately need NUMBER GO UP and FAST. One wrong turn in the STONKS market and BTC is buttered TOAST. Then no one is discussing spam on the sh*tcan of the so called "Cryptocurrency Industry." formerly known as BTC

• 
  I-AM-RAAJ🐦 (@raaj_hydb) reported 38 minutes ago

  @AmazonHelp why fast forward option is so bizarre on @PrimeVideoIN app on my android Google tv? It takes more than 10 secs for a 10 sec ⏩ My internet is fine All other ott apps I have not faced this issue On Netflix hotstar z5 sonyliv etc are very much fine. Observation:3months

• 
  Violet (@Violet_Okami) reported 38 minutes ago

  Forget the AI bubble, pop the bottom bubble. The market is overweight and overbought. Short all the amazon thigh high shops, call Michael Burry or some **** and pay him to write another article about how the market is going to crash for the 125th time. We gotta fix this.

• 
  Sumon Maiti (@Sumon_Maiti) reported 40 minutes ago

  @AmazonHelp Seller is rejecting to take responsibility worst experience with @amazonIN @amazon Totally waste of money don't purchase Amazon Basics product. I am facing problem and no one taking responsibility. 😡😡😡

• 
  Moan Joker (@DrJokerMoan) reported 41 minutes ago

  Here is the return information provided by Amazon for my top selling ASIN in the "insights and opportunities" and "voice of the customer" screens. Clear as mud. Good CX health, at risk, various short and long term return rates, and an entirely different actual customer initiated return rate. I'm either doing exceptionally well or about to get a frequently returned badge depending on which of these numbers Amazon decides is "truth" at any given moment. Can we fix this? @amznsellerhelp @AmazonASGTG @michaelpatron0

• 
  Anshul Goyal (@AnshulRGoyal) reported 48 minutes ago

  @AmazonHelp @amazonIN As you can see in the screenshot, it says that "it was rejected by me" which is not the case. And it had reached the final delivery location in Surat so shipping is also not an issue here.

• 
  Joe McWopSki (@LakesFirearmsTr) reported 48 minutes ago

  @ShadesOfPunky @the_papa_lobe @DschlopesIsBack I've heard of it, but I'm also over streaming services. I've cut it down to Netflix and Amazon Prime. I'll stick with physical discs whenever possible

• 
  Nikhil Pathak O(negative) (@pathaknikhil) reported 49 minutes ago

  @AmazonHelp Even this link is not working.. Enough time given.... Kindly do resolution.. don't buy time as has been done by your team from past 7 days

• 
  COOL CART $$ (@MANILA_147) reported 49 minutes ago

  @asvpkc @SarautaMaryam @fkeyamo Empty brain wetin concern Amazon card and this issues now illiterate

• 
  Preetika (@PreetikaGhawri) reported 52 minutes ago

  @AmazonHelp I have been waiting for 3 weeks now. Despite following up and being asked to wait, my order has been cancelled yet again due to an issue at the delivery hub. This is extremely frustrating. Please look into this immediately and ensure it doesn't happen again."

• 
  Hintze (@Hintzeboy5) reported 54 minutes ago

  @Picarat @BiancoDavinci Yeah, they know that, but I think they are doing it so they stopped deforestation in Brazil. Brazil has been cutting down the Amazon rainforest.

• 
  Hellbininfo🪽💙 (@hellaversenews) reported 55 minutes ago

  This is genuinely so weird... what is Amazon doing??? This issue seems way bigger than we think and it's so confusing

• 
  Sherry Sadie (@SadieSherr45406) reported 55 minutes ago

  @radleftreceipts The tweet highlights a real tension in tech (US worker protections vs. global talent/visa programs amid restructurings), but it simplifies a messy business reality: Xbox has real profitability issues after heavy investments, and Microsoft (like peers) manages headcount while hiring for future skills. Sharma, an American executive with a non-gaming background, is executing a "reset" for growth by 2027. Microsoft's response: Layoffs were for restructuring an unprofitable Xbox business — not to replace US workers with visa holders. H-1B employees make up a small % of the workforce overall (not Xbox-specific), and some were also impacted. Decisions are "based on business need, not visa status."gamespot.comContext on the Broader DebateTech layoffs (including Microsoft's prior rounds) often coincide with ongoing hiring in growth areas like AI/cloud. This is common across the industry (e.g., Amazon, Google). H-1B is controversial: Critics argue it enables lower-cost foreign labor and displaces Americans (especially in downturns); supporters say it fills specialized shortages. Data shows approvals for high-wage roles, but abuse concerns exist. No public evidence ties these specific Xbox cuts one-to-one to new H-1B hires. Sharma's memo focused on cost structure, not immigration or offshoring.

• 
  Jeremy Standiford (@jeremystandifor) reported 56 minutes ago

  @MsVeilMoney Subscription price is up and quality down. Increased competition from Amazon, Disney, etc. Netflix has no solution.

• 
  Vishal Gupta (@vishalcloudguy) reported 58 minutes ago

  @AmazonHelp Link is not working, give me new link

• 
  Millie Marconi (@MillieMarconnni) reported 1 hour ago

  A Stanford researcher named Fei-Fei Li once hired 49,000 strangers from 167 countries to look at pictures and answer one simple question, and their answers became the foundation modern AI is built on today. Back in 2006, Li was a freshly minted PhD from Caltech starting her first job as an assistant professor. Everyone around her was chasing the same idea, that the path to smarter AI ran through smarter algorithms. Better math. Cleverer code. She looked at the field and decided they had the problem backwards. A psychologist had once estimated that the average person can recognize about 30,000 different kinds of objects on sight. A dog. A very specific species of fern. Li's question was simple and almost embarrassing in hindsight. If a computer was ever going to see the world the way a person does, shouldn't it first be shown what the world actually looks like, at real scale, in all its variety? Nobody was building that. So she decided to build it herself. She moved to Princeton and pulled in a lexical database built by linguists there called WordNet, roughly 22,000 categories of nouns organized by meaning. That became the skeleton. Then came the impossible part. She needed millions of real photographs sorted correctly into every one of those categories, and there was no way she or her small team could label them all by hand. Working nonstop, one image a minute, with no sleep and no food, it would have taken one person almost 23 years. So she turned to Amazon Mechanical Turk, a website where you can pay strangers online to do small tasks for a few cents each. Over the next three years, those 49,000 workers in 167 countries looked at photograph after photograph and answered the same simple question, over and over. Does this picture show a dog. Does this picture show a fire truck. Each image got checked by multiple workers before it counted. By the time they were done, they had sorted more than 14 million images into over 20,000 categories. They called it ImageNet. When she finally brought it to the biggest computer vision conference in the world in 2009, the field shrugged. A hundred times bigger than any dataset that existed, and almost nobody cared. They didn't even give her a stage. They gave her a folding table in the corner of a convention center in Miami, wedged between a few posters nobody was reading. So Li did something clever. Instead of just publishing the data and hoping someone noticed, she turned it into a competition. Every year, teams would submit AI systems and see whose could recognize images most accurately against ImageNet. She built the incentive that the data alone could not create. For two years, nothing dramatic happened. Then in 2012, a team out of Toronto entered a neural network called AlexNet, trained on two ordinary gaming graphics cards. Every serious entry before it had an error rate hovering around 26 percent. AlexNet came in at 15.3 percent. An 11 point jump in a single year, inside a competition that used to inch forward by fractions of a point. That one result is the actual starting gun for the AI you use today. Not a lab announcement. Not a keynote. Just 49,000 strangers on the internet answering questions about photographs, three years before anyone realized what they had built.