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.

• Errors (47%)
• Website Down (33%)
• Sign in (19%)

Live Outage Map

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

City	Problem Type	Report Time
Fitchburg	Website Down	3 hours ago
Inglewood	Website Down	3 hours ago
Owosso	Website Down	4 hours ago
Orlando	Errors	5 hours ago
Township of Evan	Website Down	7 hours ago
‘Ewa Beach	Sign in	19 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:

• 
  Ky $hiesty 🎲🥷🏿 (@EastSideKeith_) reported 4 minutes ago

  ****** be signing 360 contracts to work security at Amazon 😭that **** is such a terrible company

• 
  brood (@broodovermind) reported 5 minutes ago

  She is saying they cannot do anything if high schoolers burned down every amazon warehouse and every wallmart

• 
  emil (@TheEmilJay) reported 9 minutes ago

  @BradyFightTalk i have a couple more value boxes coming from amazon and i'm going to try to get at least one more mega when i get home from the collective if you want to complete that set i'm down to move the ones i have. either sale or trade whatever

• 
  nancy (@veilofbeing) reported 13 minutes ago

  wait is this taxed? i’m gonna have to check my paystub bc i think it was just included in my paycheck last time which means i was taxed twice. not sure how to avoid the amazon locker issue though 😒

• 
  MidNightCodeX (@MidnightCodex0) reported 16 minutes ago

  @chamath The internet had the same sentiment problem in 2000. People hated it. “It’s killing bookstores, killing jobs, it’s a bubble.” Then Amazon, Google, and Facebook quietly built trillion dollar empires while the public was still angry. Public mood is a lagging indicator. The builders don’t wait for permission. 67% of people aren’t using AI at work yet — that’s not a failure. That’s the size of the opportunity. The companies that figure out how to make AI invisible and useful will win. The ones making it loud and threatening will fumble it exactly like Chamath is warning. The race isn’t to build the best AI. It’s to build the AI nobody realizes they’re using.

• 
  emma (@emmap72002) reported 17 minutes ago

  @loudouncats So why are their kennels so small? Did they try having them together or are they assuming there's going to be a safetyn issue? You can get cameras on Amazon for a couple of quid to monitor them 24/7. Surely they should have a couple of kennels big enough for a bonded pair.

• 
  Dr. Ether PhD (@CulverVist60210) reported 18 minutes ago

  @ScammerPayback Please blast, take down, expose and analihate (480) 618-3051 Amazon spam calls

• 
  till the levee breaks (@rocicrew) reported 20 minutes ago

  idk i feel like it’s dishonest to blame the public for a league of their own when the show was talked about a lot while airing, everyone was watching it. the issue was amazon prime

• 
  Abin (@designedbyabin) reported 22 minutes ago

  @jablamsky Microsoft’s OpenAI partnership made its Stock go Skyrocketing. Now that Anthropic is having wider success in the Enterprise, Microsoft themselves choosing Claude models for M365, Amazon inking partnership with OpenAI for RESTFUL API Agents and the fact that Google & Amazon invested more in Anthropic is hurting Microsoft’s Stock. So every announcement Google or Amazon makes about Anthropic helps their stocks. Microsoft can’t do the same with OpenAI anymore since their relationship is strained and also the market doesn’t view OpenAI favourably anymore as Anthropic and Google models are gaining more attention nowadays. Microsoft also doesn’t have Frontier models like Google now - according to MSFT the plan is to develop them by 2027. Also Google and Amazon is using their own Chips in their data centres more than Microsoft does. Every announcement by Google talking about Anthropic is driving Microsoft’s stock down. They’re returning the favour to Microsoft for what Microsoft and OpenAI did to them back in 2023 & 2024.

• 
  Jeremy Snyder (@jsnyder252) reported 23 minutes ago

  @TalkinBaseball_ @shea_station @amazon What about Pete or does that not fit the narrative since he's off to a slow start? It's a shame the seasons already over 16 games in. Never seen a team overcome being two games under .500 in April

• 
  Rambler | X -Amazonian | Account Health Expert | (@Rambler_Amz) reported 24 minutes ago

  Even when the seller actually made the mistake, the right strategy still works. Amazon does not need you to be perfect. They need you to prove that your account is now a lower risk than it was before. That is the entire game. Root cause. Fix. Prevention. Strategy over emotion.

• 
  Saikrishna (@krishsai03) reported 27 minutes ago

  @AmazonHelp @AmazonHelp issue not resolved. @amazon

• 
  Mark (@SaltWater651) reported 30 minutes ago

  @AmazonHelp No thanks. I've already talked to a "shipping supervisor" through your call me feature. my options basically boil down are "deal with it" or basically figure it out myself. So like I said.. My sourcing will be more complicated, but I won't have to deal with your companies ******* stupidity. I just have to go direct.

• 
  Pittsburgh_Martini (@JohnMartini1) reported 33 minutes ago

  @JeffBezos @amazon @FlexDriverAssoc Sir I need your help resolving an issue regarding the wrongful deactivation of my Amazon flex account based on the grounds that warehouse staff violated ToS and it was counted as strikes against me. I can provide more details if requested

• 
  Paul Gloriod (@PaulGloriod) reported 34 minutes ago

  @unusual_whales Noting like compassionate management. Fix it Amazon!

• 
  CHR¡s! (@ChrisDebate) reported 34 minutes ago

  @TheSketchyKori It is adequate and that’s the problem. Amazon has all the dough in the world

• 
  Just little old me (@Justlittle39569) reported 35 minutes ago

  @FawnsOnly @UnderYourTree Oh heck, here’s the thing, I went on Amazon & sent some flooring, thought it was something nobody would send …… I let you down mate 😖

• 
  Caroline Jelen (@CarolineJelen10) reported 40 minutes ago

  @doc_1029 Central HVAC? It may be the run capacitor. Most likely part to go. Cheap fix, maybe $15-18 off Amazon, install takes about 10-15 minutes. Simple. I've replaced one & I'm a chick. You got this! If you can't wait, get window unit for bedroom temporarily. Hard rain doesn't last

• 
  Vicky Vicky (@VickyVicky47600) reported 41 minutes ago

  @AmazonHelp @amazon @AmazonUK Hello Ruby, I have already followed all the instructions mentioned and have contacted the support team . However, the issue is still not resolved Because of this unresolved issue, my previous application was cancelled, and now the same problem is happening again.

• 
  The Sincere VP (@thesincerevp) reported 42 minutes ago

  I am a senior security engineer at one of the twelve companies that signed onto Project Glasswing. I've spent the last three weeks running Claude Mythos Preview against our production codebase. I need to tell you what I saw. On April 7th, Anthropic quietly assembled Amazon, Apple, Cisco, CrowdStrike, Google, JPMorganChase, Microsoft, NVIDIA, Palo Alto Networks, Broadcom, and the Linux Foundation into a room and told them something that changed the conversation. Their new model — Mythos Preview, unreleased to the public — had found thousands of zero-day vulnerabilities in every major operating system and every major web browser. Autonomously. Without human guidance. Including a 27-year-old bug in OpenBSD, an operating system whose entire reputation is built on being unhackable. Let me put that in context. OpenBSD's website literally says "Only two remote holes in the default install, in a heck of a long time!" That bug survived 27 years of the most paranoid security review process in the industry. Mythos found it in hours. But here's the part that made the room go quiet. They showed us what happened with Firefox. A few weeks earlier, they'd pointed Opus 4.6 — their previous model, not even Mythos — at Mozilla's JavaScript engine. Twenty minutes in, it found its first Use After Free. By the time the team finished validating that one bug and filed it in Bugzilla, Claude had already found fifty more. They ended up submitting 112 unique reports. Mozilla assigned 14 as high-severity — nearly a fifth of all high-severity Firefox vulnerabilities remediated in all of 2025. From one model. In two weeks. Then they showed us the Mythos numbers. Opus 4.6 could find vulnerabilities reasonably well. But when they asked it to actually write exploits — to turn those bugs into working attacks — it succeeded twice out of several hundred attempts. A 0.5% rate. Concerning but manageable. Mythos Preview hit 181 successful exploits on the same Firefox JavaScript engine bugs. Plus 29 more where it achieved register control. That's not a 0.5% success rate anymore. That's the model independently chaining vulnerabilities, writing JIT heap sprays, escaping browser sandboxes, and constructing multi-packet ROP chains. One of Anthropic's engineers — no formal security training — asked Mythos to find remote code execution vulnerabilities overnight. Went to bed. Woke up to a complete, working exploit. So on April 10th, three days after the Glasswing announcement, Anthropic published the blog post that's been circulating in security circles all weekend. "Preparing Your Security Program for AI-Accelerated Offense." It reads like a corporate best-practices document. Patch faster. Scan dependencies. Adopt zero trust. Design for breach. But if you read it closely, there's a sentence buried in the middle that says everything: "Mitigations whose value comes from friction — making an attack tedious — rather than a hard barrier are much less effective against an adversary that can grind through those tedious steps." That sentence just deprecated about 40% of the security industry. Rate limiting. CAPTCHAs. Non-standard ports. Extra login steps. Complexity-based deterrence. The entire philosophy of "make it annoying enough that attackers move on to easier targets" stops working when the attacker doesn't get annoyed. When the attacker is a model that will attempt the same exploit chain ten thousand times at zero marginal cost while your SOC team is eating lunch. Anthropic committed $100 million in Mythos Preview credits for defensive scanning, plus $4 million to open-source security organizations. That sounds generous until you calculate that global cybercrime costs roughly $500 billion a year, and the company publicly stated that models of similar capability will be "widely available within 24 months." So the company preparing the biggest AI IPO in history just told twelve of the largest technology companies on earth that their new model can autonomously write browser exploits, crack open operating systems that have been hardened for three decades, and that equivalent capabilities will be commoditized within two years. Then they published a checklist. I've been in security for sixteen years. I've read a lot of vendor advisories. I've never read one where the vendor was simultaneously the threat, the detector, the consultant, and the only entity offering a solution — all while preparing to go public. Anthropic built the sword, built the shield, sold the shield to the people most threatened by the sword, and released a blog post telling everyone else to patch faster. The twelve companies in that room are now scanning their codebases with Mythos. The rest of the industry is reading a five-minute blog post and hoping the checklist is enough. This is a fictional narrator. The numbers are real.

• 
  John Stempin (@WAGONStempin) reported 43 minutes ago

  @Darbybailey I sold paint/hardware/lawn goods for Sears in the 1980s when we were a juggernaut. Still one of my most favorite jobs. They made one gigantic error. They closed the mail order catalogue department the same year Amazon incorporated. Everyone is shopping at malls now, they said, no one will buy mail order. They should have been Amazon.

• 
  Drew Margulis (@dmarge18) reported 45 minutes ago

  @thenewsseeker23 @Cernovich Problem with us is our dryer vent is about 20 feet to the outside vent. The Amazon kits aren't sturdy enough. Well worth paying the professional to do it tho. Terrible idea to design the vent that long

• 
  ruiner down (@RuinerDown) reported 46 minutes ago

  @LeadingReport Im a carrier who does work for amazon in my truck and I can sadly totally see this going down. The people in these places are more drones than humans by how they're told to work

• 
  INCITE AI (@Incite_corp) reported 46 minutes ago

  Amazon’s LEO antenna unlocks in‑flight broadband scale for AWS and Prime. Amazon (AMZN) NASDAQ gains an aviation foothold as its single‑unit LEO antenna promises 1 Gbps down and 400 Mbps up. This matters because one‑day installs turn airline fleets into quick wins for connectivity revenue. The stock’s recent multi‑session rise is smaller than this new service scope over the same days, so installation speed and cabin‑wide capacity remain only partially reflected in price. Kuiper hardware now targets airlines. AWS edge services follow planes worldwide. Prime Video in flight becomes native distribution. Airline contracts set the revenue ramp path next.

• 
  Mkhabela (@IamTheTozzy) reported 55 minutes ago

  @glamfika Amazon, Walmart, Microsoft, Oracle and a host of other American, European and Asian companies operate in SA and they had no problem following the countries laws,how is this not similar to how America forced the sale of TikTok.

• 
  RDS (@Yuvi_076) reported 56 minutes ago

  @AmazonHelp @amazonIN Service is going down day by day calling executives is talking very rudely His name is Raj he is full of attitude I want to cancel my order I don't want to buy from Amazon

• 
  Links (@LordOfTheYips) reported 56 minutes ago

  @julianbanks1978 @ostentration The problem is Amazon is tired of building out “safe spaces” and getting their HR bogged down with complaints of work life balance. They’re sick of hiring Americans.

• 
  Black Edge (@BlackEdgeFund) reported 58 minutes ago

  Intel is 8 days into what could be its first 9-day winning streak in decades. Shares are up from $19.73 to $65.14 The rally started when Intel landed a deal to make custom chips for Amazon Web Services. Then came Google. Then Elon Musk's xAI. Three hyperscalers betting on Intel foundry services in two weeks. This is the same company that was trading at $40 just last month — down 60% for the year. Either Intel just found its footing in the AI chip wars, or we're watching the mother of all short squeezes before reality sets in.

• 
  NotJust10is (@Del_10is_tweets) reported 58 minutes ago

  @unusual_whales If this is true...its one of the BILLION reasons why I DO NOT deal with Amazon. How in the hell do u keep working with your DECEASED coworker laying on the floor whether near you, or 5 floors away. Fine the dirtbags , shut them down. Move them to an island somewhere.🚫👎🏽BYE!

• 
  villanelle ✨ (@elleloveCHI) reported 1 hour ago

  @EddiebroRon Do you remember The Jungle by Upton Sinclair, or Out of This Furnace by Tom Bell? To think we are returning to a time when there were no labor rights for corporate gains. It's terrible! I know everybody loves their quick shipping Amazon but somethings gotta give