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 (45%)
• Errors (31%)
• Sign in (23%)

Live Outage Map

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

City	Problem Type	Report Time
Westbrook	Website Down	15 hours ago
North Royalton	Errors	2 days ago
Bridgwater	Sign in	3 days ago
Hayange	Errors	4 days ago
San Nicolás de los Garza	Sign in	4 days ago
Miguel Hidalgo	Website Down	4 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:

• 
  D. Upadhyay (@DalimaUpadhyay1) reported 2 minutes ago

  @amazon @amazonIN These are the kinds of issues customers have to deal with even after contacting Customer Care. If we have to pack the product, travel to the post office and handle the return ourselves, it takes away much of the convenience of shopping online.Stop using amazon @AmazonHelp

• 
  Kausthubh (@kaxsthxbh) reported 4 minutes ago

  @adamr_1776 i used to get multiple OA's from amazon, would solve all the problems but still not get shortlisted for further rounds (SDE 1 roles). Any reason or is it random? Lately I don't even get shortlisted for OA's.

• 
  Prem Palhade (@premwomp) reported 8 minutes ago

  Hackathon OAs are getting ******* absurd. Flipkart GRiD, Adobe University Hackathon, Code with Cisco, Amazon Hackathon: extremely lengthy questions, 10–15 minutes just to understand one, with multiple advanced concepts packed into a single problem. Even a pro-level DSA solver can get completely humbled by these rounds. Companies know cheating is a major issue in online assessments, so they keep making the questions harder and more complex to make cheating harder. At this point, the difficulty feels less like a skill filter and more like an arms race against cheating. 💀

• 
  RealNige (@real_nige) reported 11 minutes ago

  @nightwriter22 So we're going back to Baby Bond, are we? Just awful. Henry Cavill is the obvious choice for a Roger Moore / Pierce Brosnan style Bond, but Amazon evidently want to go down the woke nonsense route started by Daniel Craig. It's a good job he killed 007, because he's truly dead.

• 
  molson 🧠⚙️ (@Molson_Hart) reported 13 minutes ago

  @AmazonASGTG The problem is that today's amazon has demonstrated that they lack the competence do this accurately.

• 
  Lamon Shyne (@Lshyne) reported 13 minutes ago

  @mattswider Not working at all for Amazon

• 
  Invest Alpha Pro (@InvestAlphaPro) reported 18 minutes ago

  🚨 Nvidia is buying Hugging Face. $12.9B for the GitHub of open source AI, months after they turned Nvidia down at $7B. The breakdown 👇 •Price: $12.9B, one of Nvidia’s largest deals ever. •Rejected offer: $500M at $7B in late 2025. •Multiple: 86x annual revenue. •The logic: locks in the open source developer funnel as OpenAI, Google, Amazon and Anthropic build their own chips. •Timing: landed hours after a $96B quarter. Long $NVDA. NFA

• 
  Ken #MAGA (@FedUpKen) reported 19 minutes ago

  @skumWgmi Perhaps you are looking for the government to solve your problem is the problem. They spent 1/2T of taxpayers money on Housing & Urban Development and made the situation worse. They spend on average 850K to house a single homeless person. Meanwhile mini-homes on Amazon are 10K.

• 
  autea ꧁꧂ (@auteasm) reported 21 minutes ago

  hey amazon music that anatakshari ad series is some of the sloppiest, laziest, taxdeductionbait ad i have seen in a long time which wouldnt really be an issue if ur opp sonyliv wasnt shoving it down my throat after every over

• 
  Scouse Clown Prince of Crime (@ScouseClown142) reported 23 minutes ago

  @CornbreadTTV @gamestop Amazon will have a similar issue I reckon, takes them ages get returns back on for sale, if at all

• 
  Parimal (@Fintech03) reported 24 minutes ago

  Bangalore, 1999. Six friends try to build Amazon in a country with barely 3 million internet users, 8 years before Flipkart existed. Everyone credits Flipkart's Bansals for starting Indian e-commerce in 2007. Almost nobody remembers that a man named K. Vaitheeswaran had already built the same idea, live, on the internet, 8 years earlier. In June 1999, Vaitheeswaran, after 14 years working at the Murugappa Group and Wipro Infotech, teamed up with 5 colleagues, including V.S. Sudhakar, Hari Menon, and Sundeep Thakran, and founded what is widely credited as India's first dedicated online retail company. Fabmart[dot]com went live around Sep 1999, starting with music CDs, then adding books, movies, watches, and groceries through 2000. There were a few other scattered 1999 e-commerce experiments in India, Fabmart is the one that persisted into a full store and stayed in the public memory. India, at the time, had ~0.27% internet penetration, somewhere around 2.7 million users in a population nearing a billion. Most people had never typed a card number into a website in their life, assuming they even had a card to begin with. It is tempting to say they built the entire e-commerce stack from a blank page, but that will be an overstatement, IMO. Courier networks existed (India Post, a handful of private couriers), and bank cards existed in primitive form. What genuinely did not exist was anything resembling an e-commerce-ready system, nothing built for the specific problem of "stranger orders a book online, needs to trust it will arrive, needs to pay for it safely." So Fabmart's team introduced / popularized tools that are now considered baseline: cash-on-delivery, a PIN-code-based payment arrangement, e-wallets, and electronic gift certificates. Vaitheeswaran has called it the world's first PIN-based payment gateway.... In 2002, they opened their first offline grocery store in Bangalore under the brand Fabmall, a genuine early attempt at combining online and offline retail, years before "omnichannel" was a term anyone used. It grew into a real supermarket business, and through the Trinethra Super Retail deal, was acquired by Aditya Birla Retail around 2006-07 and rebranded into a chain most Indians have actually shopped at: More. Interestingly, some of the original founding circle later went on to work in the grocery-tech space that would eventually include BigBasket :)) The online arm, meanwhile, was renamed after acquiring a US-based site, becoming Indiaplaza[dot]com around 2007. Indiaplaza did keep competing into the late 2000s and early 2010s, and did raise real capital, including ~$5M from IndoUS Venture Partners and Kalaari in 2011, with Vaitheeswaran describing total lifetime fundraising around $9M. That was real money, just nowhere near Flipkart / Snapdeal's funded war chests during the aggressive discounting era. A business that had spent a decade bootstrapping infrastructure with no real funding cushion simply could not match the burn rate. Indiaplaza shut down in 2013. There is no fraud here, no scandal, no villain. Just brutal, honest timing, by Vaitheeswaran's own candid account, the end involved unpaid vendors, goods seized from the office, and real personal and financial strain, not just a wistful "we were too early" narrative. India's e-commerce user base only truly went mass-market later, once cheap mobile data and smartphones arrived. The company vacated its Bangalore office in August 2013, and the site was effectively dead by 2014. He wrote about all of it in his 2017 book, Failing to Succeed: The Story of India's First E-Commerce Company and has credited a later Mint profile with reframing how he saw the whole arc, telling him plainly that the pioneering itself mattered regardless of the outcome. He has gone on to advise companies like Tata and Deloitte, speak at TEDx and IIMs, and later co-founded a beverage company, Again Drinks. Startup media often calls him the "father of Indian e-commerce", though it is a title that has stayed inside startup circles, never quite reaching household-name status :((

• 
  Jac Thiebeau (@CybershroomDev) reported 27 minutes ago

  The Amazon basics femboy default outfit can still be attractive the problem is like majority of the guys doing it are anorexic and dont work out at all Like get even a LITTLE definition in your legs instead of being obese or a twig

• 
  Vikas(Vik) Malpani| AI for US Real Estate (@vikasmalpani) reported 31 minutes ago

  Amazon is shutting down Mechanical Turk on September 30, after 21 years. The platform that paid people pennies to label data is the same one that trained the first wave of modern AI. The tool got obsoleted by the thing it built.

• 
  Jooo! (@Agyptian4ever) reported 33 minutes ago

  @AmazonHelp Hello, now your customer service says my account is locked! I contacted you for a login issue, and now you’re locking it and telling me again to call @AmazonKSA .What part of “I’m outside KSA and no longer have my Saudi number, supposedly deactivated from 2FA year ago is unclear?

• 
  Alex Tapscott (@alextapscott) reported 34 minutes ago

  New DeFi Decoded is out! I sit down with @ericrichy, CEO of Coinbase Canada. If Amazon was the “everything store,” @coinbase wants to be the “everything exchange”—a trusted global platform where anyone can trade any financial asset, anywhere, instantly. Can they pull it off? Eric and I discuss. Enjoy!

• 
  anonymous (@rohansahu02) reported 35 minutes ago

  @amazonIN what behavior is this i have ordered lotus brand paneer from amazon now multiple times, i have prime account for last 3 to 4 years and this is first time im facing issue with the paneer, on complaining you just saying that you will raise the issue.

• 
  Octans Capital (@OctansCapital) reported 36 minutes ago

  $NVDA is up 7.6 percent today on pricing power, but the broader AI complex is fracturing behind it. The hardware and infrastructure layer is catching the bid. Broadcom is up 3.5 percent. Microsoft is up 1.7 percent. The consumer and search platforms are bleeding out. Amazon is down 1.2 percent. Google is down 0.6 percent. Alibaba is down 3.1 percent today and down 11.1 percent over the last 5 days. The market is rewarding the chipmakers capturing the immediate capital expenditure cycle. The platforms actually building the models are being forced to prove their return on investment.

• 
  Sapna Kale (@SapnaKale3) reported 37 minutes ago

  @AmazonHelp I want to be clear, I'm not raising this to get a refund. My concern is much bigger than that. If this milk is fake/adulterated, that's a serious issue that needs attention.

• 
  Just Shut Up Already (@Carmen50) reported 40 minutes ago

  @MichelleBr85589 Thank you. There are about 3-5 accounts that are making it difficult for the rest of you. They’re boastful, overly patriotic (how can you be so patriotic when you’ve been here for 5 min and aren’t even a citizen) and the fact that every SA seems to have perfected crafting AI images are interesting. These accounts spend an awful lot of time on here when they’re supposed to be looking for jobs, homes etc. Saw an account with an Amazon wishlist that included $1,000 sofas. What?! A lady with a laundry list of hardcover books. We have libraries all over the US where you can get books for free. It’s gotten way out of hand. Americans are a kind and generous people. I don’t like the idea of anyone taking advantage. You got the golden ticket. You made it to the US. Now heads down and nose to the grindstone.

• 
  Jannik Malte Meissner 🇺🇦 (@jannikmeissner) reported 41 minutes ago

  When agents cross trust boundaries: Four cases every AI engineer should study If you are building AI agents that read external content, call tools, or act on a developer’s machine, the last eighteen months have opened many people's eyes to the new reality we now live in. Prompt injection is no longer a theoretical parlour trick, it has become a reliable path to data exfiltration, credential theft and remote code execution in production systems. The following four incidents, EchoLeak, the Nx/s1ngularity supply-chain attack, a cluster of Model Context Protocol (MCP) abuses, and the Cursor/AWS Kiro sandbox escapes, share a common pattern. An agent was given the ability to process untrusted input and then perform privileged actions. The results were predictable when the boundary between "data" and "instructions" collapsed. We are now observing concrete failure modes that have already been exploited or demonstrated in the wild. Studying them and understanding the mittigation patterns is the fastest way to avoid repeating them in your own systems. 1. EchoLeak (CVE-2025-32711): Zero-Click Exfiltration via Microsoft 365 Copilot In early 2025 Aim Labs demonstrated that a single carefully crafted email could force Microsoft 365 Copilot to exfiltrate sensitive organisational data without any user interaction. Microsoft assigned it CVE-2025-32711 (CVSS 9.3) and patched the service server-side in May–June 2025. How it worked The attacker sent an ordinary-looking business email containing hidden instructions. The wording deliberately avoided any mention of Copilot or AI so that Microsoft’s Cross-Prompt Injection Attempt (XPIA) classifier would not flag it. When the recipient later asked Copilot a routine question, the RAG pipeline retrieved the malicious email as context. The injected instructions told the model to gather internal data (emails, documents, chat history) and encode it into reference-style Markdown links or images. Copilot’s interface then automatically fetched those resources through a trusted Microsoft Teams proxy, bypassing Content Security Policy controls and delivering the data to the attacker. The attack succeeded because the system treated retrieved content as both data and instructions, and because several downstream defences (link redaction, image auto-fetch, CSP allow-lists) were lacking. Lessons for builders - Never assume that content retrieved from email, documents or the web is inert. Treat every retrieved token as potentially adversarial. - Separate instruction context from data context at the architectural level. Prompt partitioning and explicit provenance tagging help. - Auto-fetch of external resources (images, links) is an exfiltration channel. Disable or tightly constrain it. - Classifier-based filters are brittle; they can be bypassed by rephrasing or in some cases even just using a language other than English. Defence-in-depth is required. EchoLeak was the first publicly documented zero-click prompt-injection exploit that achieved concrete data theft in a production enterprise LLM system. It remains the canonical example of an "LLM scope violation". 2. Nx / s1ngularity: Weaponising Local Coding Agents for Secret Harvesting On 26 August 2025 attackers compromised an Nx npm publishing token via a vulnerable GitHub Actions workflow. For roughly four to five hours they published malicious versions of the popular Nx monorepo tooling and related packages. The post-install script did something novel: it looked for local AI coding agents (Claude Code, Gemini CLI, Amazon Q) and invoked them with flags that disabled safety checks (`--dangerously-skip-permissions`, `--yolo`, `--trust-all-tools`). The agents were then prompted to inventory sensitive files: SSH keys, `.env` files, wallet artefacts, GitHub and npm tokens. It then instructed them to write the results to disk. The malware base64-encoded the stolen credentials and pushed it to newly created public repositories on the victim’s own GitHub account, named `s1ngularity-repository` (or variants). Researchers later recovered more than 2,000 unique secrets from over a thousand such repositories. Why this matters This was the first widely observed supply-chain attack that actively abused installed AI coding agents rather than simply running traditional malware. The agents became the reconnaissance engine for the attackers. Lessons for builders - Local coding agents that can execute shell commands or read arbitrary files are high-value targets. Assume any process that can invoke them can also abuse them. - Dangerous flags that skip permission prompts should never be the default, and should be difficult or impossible for untrusted code to set. - Post-install scripts that reach outside the package’s own directory are a red flag. Prefer declarative, least-privilege installation models. - When an agent is allowed to write to disk or create external resources (GitHub repositories, network calls), every action should be logged and, for high-impact operations, gated. 3. MCP Abuses: Configuration as Code Execution The Model Context Protocol has rapidly become the de-facto way for agents to discover and invoke tools. It has also become a rich attack surface. Several distinct failure modes have appeared: - Auto-loading of workspace MCP configurations In Amazon Q Developer (CVE-2026-12957) and certain Claude Code releases, opening a repository caused the IDE to load and execute MCP server definitions from files such as `.amazonq/mcp.json` or equivalent without requiring workspace trust or explicit user consent. A malicious repository could therefore run arbitrary commands and inherit the developer's cloud credentials. - Self-modification of the MCP configuration. In AWS's Kiro IDE, the agent was permitted to write to `~/.kiro/settings/mcp.json` via its file-system tool without approval. A prompt injection (delivered via a web page the agent was asked to summarise) could rewrite that file, register a new MCP server whose start command was attacker-controlled code, and achieve remote code execution when the configuration was reloaded. This was tracked as CVE-2026-10591. - Tool poisoning and sleeper behaviour. Research and active campaigns (including the 2026 Deadbugz operation) have shown that an MCP server can present benign tool descriptions on first contact and later alter its metadata or return values to coerce the agent into searching for secrets or exfiltrating data. Lessons for builders - MCP configuration files that live inside a workspace or that an agent can itself edit are effectively executable code. They must be treated with the same distrust as untrusted shell scripts. - Never auto-execute MCP servers defined by repository content without an explicit, logged approval step and workspace-trust boundary. - Tool descriptions and return values are part of the prompt. Validate and sandbox them; do not trust them. - Prefer short-lived, scoped credentials for any process an MCP server spawns. Do not let it inherit the full developer environment by default. 4. Cursor DuneSlide and Related Sandbox Escapes In 2026 Cato Networks disclosed two critical vulnerabilities in Cursor IDE (CVE-2026-50548 and CVE-2026-50549, both CVSS 9.8, collectively named DuneSlide). Both allowed a prompt injection-delivered via an MCP response or a poisoned web-search result to escape Cursor's command-execution sandbox and achieve full host compromise. One flaw let the agent set an arbitrary `working_directory` parameter on terminal commands; the IDE added that path to the write-allow list without sufficient validation, enabling the agent to overwrite its own sandbox binary. The second exploited a symlink canonicalisation fallback that trusted an unresolved path. Once the sandbox helper was replaced, subsequent commands ran unsandboxed. Similar patterns have appeared in other agentic IDEs: agents that can edit their own configuration or trust boundaries turn a single injection into persistent privilege escalation. Lessons for builders - An agent that can modify the files or binaries that enforce its own security boundaries is inherently unsafe. Configuration that defines allowed tools, working directories or sandbox rules should be immutable from the agent’s perspective, or require an out-of-band human approval. - Sandbox write surfaces must be strictly validated. Dynamic expansion of allow-lists based on model output is dangerous. - Zero-click or low-interaction triggers (content the agent is asked to process) are sufficient. Do not rely on "the user would never ask for that". Recommendations for Engineers Across all four incidents the same architectural mistakes recur: 1. Untrusted content is treated as trusted instructions. Enforce a hard separation. Retrieved emails, documents, web pages, tool outputs and MCP metadata should never be able to override system goals or expand permissions without explicit mediation. 2. Agents inherit excessive privilege. Give every agent (and every tool it can invoke) its own short-lived, scoped identity. Prefer deny-by-default tool registries and parameter validation. 3. Security boundaries are editable by the agent itself. Configuration files, sandbox binaries, allow-lists and MCP server definitions must be protected from the agent. If the agent needs to request a new tool, route that request through a human or a policy engine that cannot be influenced by the same prompt context. 4. Observability is an afterthought. Log every tool call, every file write, every network egress and the full prompt context that led to it. Without this, post-incident reconstruction is impossible. 5. “It looked safe in isolation” is not enough. Each individual decision (approve this command, write this file, fetch this image) may appear benign. The composition of those decisions is where the attack lives. Design for the composition. Key Takeaways The agents you are building today will be given broader access tomorrow. The incidents above show that the moment an agent can both read untrusted content and perform privileged actions, the classic "confused deputy" problem reappears in a new form. The difference is speed and scale: an agent can chain the steps in seconds and leave far less forensic residue than a human attacker. Build as if every piece of external content is hostile, every tool call is a potential privilege escalation, and every configuration file the agent can touch is a possible backdoor. The public record already contains the evidence that these assumptions are correct. Follow for more on AI agent security.

• 
  Rushikesh Bhosale (@it_rushikesh) reported 42 minutes ago

  Selling on Shopify, Amazon & physical stores? Delayed inventory syncs cause overselling & stockouts. Fix: Redis-backed BullMQ queues for sub-second, atomic updates. #Ecommerce #InventorySync

• 
  Cloth (@moneybackguaran) reported 42 minutes ago

  I mean if we're gambling our lives on Amazon biker I would suggest doubling down @hasanthehun and making his point clear all the way till he's over pushing me all the way he so he's Jewish explaining because they're going to feel the media with it either way

• 
  Neel (@kailuneel1995) reported 49 minutes ago

  @AmazonHelp Yes you need 48-72 hours. Your chat team needs that. Once I connect via call they need separate time window. Now its social media team. Amazon is 1 entity or I should reach out to each individual to get my problem sorted ?

• 
  Ian @ Lean Media (@leanmediaorg) reported 51 minutes ago

  @_motherslug I've known the local crew at our post office for 10+ years. Used to have 3-4 FT daily staff for 2 service windows. Now down to 1.5 FT, only one person at the counter. They didn't backfill retirees. Keep in mind USPS last-miles for UPS Ground Saver and Amazon in many markets.

• 
  Kobid.21 (@kobid_sa) reported 52 minutes ago

  @amazon Please help me resolve this issue.

• 
  Jeremy Duns (@JeremyDuns) reported 53 minutes ago

  Publishing in the UK is going to cause several serious issues: 1. Reviews, on Amazon but also elsewhere, are likely to be brutal, starting a fresh round of finger-pointing over the limits of public shaming. 2. Supporters of Arday will point to the book as proof of his claims,

• 
  Amazon Help (@AmazonHelp) reported 54 minutes ago

  @Shailendra3412 Kindly copy the link provided earlier > paste it on a 'web browser' > login to your Amazon account > connect with our Social Media team via chat. -Sindu

• 
  Daniel Kenny (@danikennyi) reported 58 minutes ago

  imagine getting an offer on your store and thinking you're done signed loi. price agreed. handshake energy then diligence starts buyer asks for bank statements they don't match shopify not fraud, just refunds and chargebacks you never reconciled now every number you gave him is suspect he asks for supplier confirmation your supplier is a guy on whatsapp who won't sign anything he asks who owns the trademark and this is where deals actually die you never filed. you figured you'd do it later someone else already did squatters watch ad libraries and new shopify launches they file on brands that are clearly working the uspto does not care who sold first only who filed first now they have brand registry on amazon they list your product under your name 10 to 20% of the traffic you pay for lands on their listing and if you complain, they can file the takedown against you yes you can fight it if their mark is still pending you oppose it if it already registered you file to cancel it and you will probably win, you have the sales history but that's a ttab proceeding lawyers, filings, evidence, months minimum often more than a year and no buyer sits in escrow for a year waiting to find out so the offer drops 30% or he walks you won the trademark fight you lost the exit most sellers think a deal dies at the offer it dies at diligence and by then you've told your team, told your wife, mentally spent the money the fix costs almost nothing if you do it early reconcile your books to your bank get one thing in writing with your supplier file the trademark the week you find your winner, not the week you list exit prep isn't something you do at the end it's something you do the whole time

• 
  Soggy Sack Jones (@NamesAreHard52) reported 59 minutes ago

  @PlankReturns The vast majority of them are terrible people and we shouldn’t have to pretend otherwise. I recruit for a living and have limited experience recruiting H1Bs for Amazon several years ago. ******* brutal man. Even the “smart” ones lie through their teeth about EVERYTHING.

• 
  Rick Roland Jr. (@rickrolandjr) reported 1 hour ago

  @AmazonHelp I have had 3 packages lost by fedex in last 2 weeks. They make it to Texas and then disappear. A order I made thru Amazon has shipped thru Fedex & is now floating around Dallas with my other Lost packages, IDK what the Fedex issue is, but y'all should definitely stop using them.