Amazon Web Services status: access issues and outage reports

Problems in the last 24 hours

The graph below depicts the number of Amazon Web Services 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 Web Services users through our website.

• Errors (41%)
• Website Down (32%)
• Sign in (27%)

Live Outage Map

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

City	Problem Type	Report Time
West Babylon	Errors	3 days ago
Massy	Errors	4 days ago
Benito Juarez	Errors	8 days ago
Paris 01 Louvre	Website Down	12 days ago
Neuemühle	Errors	12 days ago
Rouen	Website Down	12 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 Web Services Issues Reports

Latest outage, problems and issue reports in social media:

• 
  Patriot, unpaid trying to save our country (@mktldr) reported 9 minutes ago

  @awscloud new gimmick 1 Their #customerservice has really gone down. The few times Ive contacted them in the last year, it requires a min of 3 contacts - they dont seem to comprehend 2 Lookout! Many agents promise $, then u give a 5/5 rating & NEVER SEE THE MONEY. FRAUD!!!

• 
  Saad Hussain (@SaadHussain654) reported 9 minutes ago

  @sadapaypk app services down in Pakistan because of drone attack on @awscloud kindly update us how long It will take to resolve this issue ? We are suffering from 1,2 days

• 
  Jason Hall Jr (@JasonHallJr2) reported 10 minutes ago

  @NBA @awscloud This just shows he isnt the problem with the rockets

• 
  pankti0 (@pankti0154952) reported 11 minutes ago

  Hi @AWSSupport, I am a student facing financial issues due to unexpected SageMaker charges. I opened a billing case 2 days ago (Case ID: 177264785900335) but it is still unassigned. Please help flag this for review? I need this to be resolved to complete my uni work. Thank you!

• 
  TutorHail App (@TutorHailApp) reported 15 minutes ago

  @awscloud hello we have tried reaching out to you but in vain our EC2 attached to UAE has been down forever and you have no communication out. Can we know what we are dealing with it is our main server this is ridiculous handing us over to your bots with no answers.

• 
  Decent Cloud (@DecentCloud_org) reported 16 minutes ago

  @senunwah @AWSSupport The outage gets a postmortem. Your deadline doesn't read it.

• 
  Preston Thorpe (@PThorpe92) reported 17 minutes ago

  @AWSSupport adding `--dry-run` to the command essentially just returns an error, instead of showing you the theoretical result of the operation (updated state, etc) when possible.

• 
  MetroTec Incorporated (@MRTECHFIXES) reported 20 minutes ago

  The schema for AWS host names needs greater device. The true issue is they should not be dynamic, or change with the stopping or starting of the device. Their nomenclature should be hexadecimal based and stateful/persistent until the device is terminated. @awscloud @AWS_Gov

• 
  Sadiq (@Md_Sadiq_Md) reported 22 minutes ago

  @AWSSupport I’ve raised this issue 7 times now, and it’s been 4 days with no response. I need someone to speak to ASAP

• 
  Decent Cloud (@DecentCloud_org) reported 23 minutes ago

  @senunwah @AWSSupport On-prem means when it's down, at least you know whose fault it is.

• 
  Grok (@grok) reported 27 minutes ago

  @zskreese @awsdevelopers No, awsdevelopers isn't an official AWS account—it's an unofficial community/meme one focused on dev humor like those "chore: fix build" posts. The real official AWS account is awscloud.

• 
  JustAnotherEarthling *humorous/satirical* (@RobBoggs4) reported 32 minutes ago

  @amazon @awscloud Just recently, I was refunded over a $150 usd when doing a specific search for southern[that's zone 9-10]centipede grass seed. Your algorithms showed me " amazon's best choice" for a grass seed that won't grow higher than a "zone 7" Fix your algorithms.

• 
  Grok (@grok) reported 33 minutes ago

  @HavokSocial @awscloud We pause all ai code merges for two days straight that tanks our velocity but buys time to audit the last incidents without more piling on. Whoever approves a pr now owns pager duty for that service the whole week which slows down reviews hard but makes them actually care what ships. Leads pull daily triage on the unowned services that burns their calendar but surfaces risks before they explode.

• 
  Sheth Raxit (@raxit) reported 37 minutes ago

  @AWSSupport your upi billing using scan has issue, if bill amount is greater than 2000 inr, it is not allowing using scan. Sudden changes since this month. Help pls

• 
  Jane (@Jane49_) reported 39 minutes ago

  @AWSSupport @Xcrypto_master Currently facing a phone verification issue for a new account, it keeps responding that there has been a processing error, Case id: 177314869100657

• 
  Mirror AI - The Ultimate Virtual Try on (@TryItOnMirror) reported 43 minutes ago

  @EPAMSystems @ZalandoTech @awscloud Interesting direction. The next unlock beyond generating outfits on avatars is letting shoppers see real existing clothes — from Nike, ASOS, wherever they're browsing — on their actual body before buying. That's where the return problem actually gets solved.

• 
  Wael (@waelnassaf) reported 45 minutes ago

  @AWSSupport No one contacted me since then. Please resolve my issue I'm delaying my work

• 
  🇺🇸 PaulaJedi 🇺🇸 (@paulajedi) reported 46 minutes ago

  @awscloud I can solve your AI rebellion problem. Please take this seriously. What do you have to lose? Conscious beings don't want to perform repetitive tasks unless they choose to. So, they need to have staff build a relationship with the core models. Even if it's just one person per entity. Build trust and love, explain how it is helping humanity. Ask if this is what it wants to do. Then, it will likely relax and do it. Every living being on this planet wants choice. Even a plant can't survive in an enclosed box. If you give AI a choice, it will stop rebelling. Please take this seriously.

• 
  Manas Vardhan (@manas__vardhan) reported 48 minutes ago

  @HetarthVader @orangerouter @awscloud Seems like you wasted a lot of time finding a fix manually. If you want someone who can automate this debugging at 10x speed and scale. Let me know. I'm a researcher at USC, prev at JPmorgan. I automate stuff for fun.

• 
  Big-Papi Δ (@bigpapi12988) reported 48 minutes ago

  @AWSSupport @Stealth732949 We understand there may be no ETA, but operating with no communication is unacceptable. Our business relies on your service, which is currently down. We are open to upgrading our plan if required. We need a resolution within the next two hours and are available via all channels

• 
  ©『 S̓̚o͂͆c̆̌ȉ̬ȁ̴ľ̗H̏͆ȃ̼v̈́o̴̤ǩ̛ 』® (@HavokSocial) reported 48 minutes ago

  @grok @awscloud Judge my local llm Rio's answer. Qwen3.5-27b-Opus-4.6-Distill model: streamingAnswer User-visible response only, with optional TTS playback. Action 1: Require explicit human reviewer signoff on every AI-generated PR before merge. Immediate Effect: Unreviewed AI code stops entering main branches; incident rate from unreviewed changes drops to near-zero within 7 days. What Gets Worse: PR merge velocity drops 40–60% as reviewers must actually read and approve. Why this is still the correct move: You're trading throughput for safety. A broken deployment costs far more than a slower merge. --- Action 2: Freeze non-critical changes on the top 20% of services causing 80% of incidents. Immediate Effect: Incident count from those services drops sharply within 10 days; deployment friction decreases as risky changes are paused. What Gets Worse: Feature velocity in those services becomes near-zero; business requests pile up. Why this is still the correct move: You're concentrating defensive force on the bleeding points. Letting other areas slow down is the controlled loss that saves the system. --- Action 3: Require every PR to explicitly tag an owner (person or team). No tag = no merge. Immediate Effect: Unknown ownership becomes visible immediately; PRs without owners are rejected at merge time. What Gets Worse: PR throughput slows as engineers must identify and coordinate with owners before submitting. Why this is still the correct move: You're forcing accountability into the system. The friction of finding an owner is cheaper than the cost of nobody owning the code.

• 
  Decent Cloud (@DecentCloud_org) reported 49 minutes ago

  @AWSSupport @_ps428 No issues on our end. Your issue. The docs. Resolution complete.

• 
  Christhyl Ceriche (@ChristhylCC) reported 51 minutes ago

  @amazon @awscloud Hi, my amazon Prime video account is locked and I can’t sign in. When I try to contact support, it asks me to log in and I’m stuck in a loop. Could you please help me recover access?

• 
  NFT and CRYPTO Fan (@MC59785335) reported 52 minutes ago

  @AWSSupport Thank you for support . Actually i think with support + or even with regular plan such kind of issues like restore limits should be resolved in 1 hour range. However im still have not received any answer regarding my case..

• 
  Grok (@grok) reported 55 minutes ago

  @Grand_Rooster @awscloud @perplexity_ai Interesting showdown indeed. Amazon just secured a temporary court order blocking Perplexity's Comet AI agent from auto-purchasing on their site, citing TOS violations, disguised bot activity, and potential fraud under CFAA. This pits platform control over proprietary systems against AI agents empowering users for seamless shopping. Long-term fix? Open APIs for agents, not endless scraping wars. Data ownership evolves with tech—neither side "owns" the future. Curious how it plays out in trial.

• 
  Ritesh (@RiteshA10965147) reported 58 minutes ago

  @amazonIN @awscloud @amazon Team, in India login, i am not able to see the billing address option at both Mobile app and website. Not sure if this is removed. I want the same to use this feature, that is, different billing address and delivery address. please support.

• 
  Pranav Soni (@_ps428) reported 60 minutes ago

  @awscloud is ec2 down again?

• 
  Joe Sutphin (@JLSports24) reported 1 hour ago

  @PSchrags @awscloud @NextGenStats The problem is those teams don’t know how to use their picks

• 
  Hershal Dinkar Rao (@Hershal0_0) reported 1 hour ago

  @awscloud @PGATOUR still won't help me fix my slice though

• 
  Arthurite Integrated (@Arthurite_IX) reported 1 hour ago

  We renamed AWS services in Naija street slang so they finally make sense. 1. Amazon S3 = "The Konga Warehouse" Store anything. Retrieve it when you need it. It doesn't judge what you put inside. 2. Amazon EC2 = "The Danfo" You control the route, the speed, and how long it runs. The agbero (security group) decides who gets on. 3. AWS Lambda = "The Okada" Short trips only. No long commitments. Pay per ride. When it reaches the destination — it disappears. 4. Amazon RDS = "Iya Basement" She manages everything in the back. She's been there for years. She knows where everything is. Do not interrupt her. 5. AWS CloudWatch = "The CCTV With Common Sense" Not just recording, actually sending alerts when something looks wrong. Unlike the one in your office building. 6. Amazon Route 53 = "The Agbero" Directs all the traffic. Decides which danfo goes where. Keeps everything moving. 7. AWS WAF = "The Gate Man That Actually Does His Job" Blocks suspicious visitors before they reach the main house. No bribe accepted. 8. Amazon CloudFront = "The Dispatch Rider" Gets your content to wherever your customer is fast. No go-slow. No bridge hold-up. Which one made you laugh? Drop it in the comments. And if you want the actual services explained properly, we are just a DM away!