eBay status: access issues and outage reports

Problems in the last 24 hours

The graph below depicts the number of eBay 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 eBay users through our website.

• Website Down (44%)
• Sign in (38%)
• Errors (18%)

Live Outage Map

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

City	Problem Type	Report Time
Saltburn-by-the-Sea	Website Down	4 minutes ago
Berlin	Sign in	4 hours ago
Libourne	Sign in	7 hours ago
Montréal	Sign in	8 hours ago
Waldshut-Tiengen	Sign in	8 hours ago
Saltburn-by-the-Sea	Website Down	9 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.

eBay Issues Reports

Latest outage, problems and issue reports in social media:

• 
  erin🪩 (@wallowsgarden) reported 8 minutes ago

  whoever is trying to outbid me on ebay for this wallows OK vinyl, STAND TF DOWN!

• 
  Echo tone Music (@EcHoToNe) reported 10 minutes ago

  @tm1515152005 what platform? cause that makes a huge difference if its your own, like Squarespace, or a third party like eBay. But one thing if you ship alot like we do (everyday) is google every address and make sure it matches and it not a weird freight forwarding company. This mitigates chargeback issues.

• 
  Mikey J Sports (@mikeyjsports) reported 12 minutes ago

  Was Ebay down today? Or is it a North Jersey thing?

• 
  𝙧𝙮𝙖𝙣 🫧🐜 (@ryanaffect) reported 14 minutes ago

  @pokemondealsuk shame these will all go to the people that’ll just list it on ebay for 3x the price, terrible

• 
  mako (@guycalmdown) reported 23 minutes ago

  @v4mp1r3bit3 I think all those old bam adio shoes suffer from that, ive seen a ton on ebay with the same issue and they still go for a lot sadly

• 
  〠 (@JSawyN54) reported 25 minutes ago

  @GodsWrestler133 @CardPurchaser I vend but when I’m not selling at a show… I’m a marketplace guy. I know eBay is good and safe but I’ve NEVER not eventually sold something on marketplace… even the shipping on there is alright. I’d rather slow sell on FB than quick sell and get played.

• 
  Laura⁷ (@lollylix) reported 25 minutes ago

  @bumpkinsTV @flashivers some of us cannot afford trying to find new issues through amazon/ebay and paying +20€ shipping and like 10€ per comic

• 
  Ernie (BigE) 🇺🇲 (@SouthernGatorz) reported 28 minutes ago

  @eBay This is bullshit. You make people add a payment method to bid, but then you let them ghost and I have to wait 4 days? Either charge the damn card the second they win or don’t let them bid at all. Winning an auction should lock in the payment immediately. Fix this crap.

• 
  raiden (@raidenfomo) reported 31 minutes ago

  THIS WOMAN BUILT A CUSTOM SERVER FOR A CALIFORNIA COMPANY THAT KILLED THEIR $10,000 A MONTH CLAUDE API BILL. The rack on her bench is a paid order. She builds these in Los Angeles from parts data centers threw out: a used Dell chassis from a data center liquidation sale in California, two Xeons, two 24GB Tesla P40s off eBay. NVIDIA sold each card for $5,700. She pays $180. Type Tesla P40 into eBay and sort by price: $150-250, thousands of them. $1,480 in parts, Ollama for free, 20 minutes to set up. A 32B model runs inside the office and nothing the company sends leaves the building. The client's software was firing thousands of calls a day at Claude: sorting documents, tagging tickets, drafting replies. $10,000 a month in API bills. Most of those calls never needed the top model. Now everything hits the box first, and only the hard ones still go out to Claude. The bill dropped to $840. The box eats electricity and nothing else. Three companies are waiting in the queue behind this one. Your own version starts with one $180 card in a spare slot. For herself she bought a $599 Mac mini. It runs a 14B model in silence, adds $4 a month to the power bill, and covers everything she used to pay $300 a month in subscriptions for. Paid for itself on the second billing date. She builds full racks for companies. Her own AI is the size of a sandwich, and it's enough. Save this while used 24GB cards still go for $180.

• 
  rvsh (@rvsh53867562) reported 31 minutes ago

  I watched the interview 2 times. First I thought that if ebay is taking notes on these Interview, they might fix it themselves without us. Then i reminded myself they’d have to fire themselves. $GME 🎯 $EBAY @amitisinvesting @ryancohen

• 
  Big Pappa PUMP (@Xeloris1) reported 32 minutes ago

  @foxenflask If Cohen is waiting on price improvement he might be waiting awhile or even risk eBay going over the 125 bid price. It's unfortunate I think it's a timing problem, because I think most of this draw down was external forces of sector relation. Ai has pulled away a lot of cap

• 
  NO1ennn (@N01ennn) reported 34 minutes ago

  A HOMELAB BUILDER JUST PAIRED AN AMD EPYC 77 CHIP WITH 512GB OF DDR5 AND LLAMA.CPP TO RUN DEEPSEEK V3 ENTIRELY FROM RAM WITHOUT A SINGLE GPU 00:03 "my very own AI super computer, not exactly, but it should be quite capable. the goal is to run large LLMs entirely from RAM with a decent context window" the stack is minimalist. one AMD EPYC 77 series processor, 8 memory channels, 512GB of DDR5 ECC, llama.cpp compiled for CPU inference. no GPU. no CUDA. no waiting for a used 3090 to appear on eBay running DeepSeek V3 and Llama 4 from system memory means inference happens at CPU speeds, slower per token but unlocks model sizes that would need a $40,000 8xH100 server to fit on GPU. 405B parameters loaded straight into RAM the article ranks local AI compute from $180 Tesla P40 up to $4,199 Mac Studio. this build sits in a different lane entirely. not GPU acceleration, but raw memory capacity. RAM is the new frontier for anything above 200B parameters most people wait out the GPU shortage. he built an inference machine that never needed a GPU in the first place save this before every homelab realizes their server chassis was already an LLM machine waiting for enough RAM ↓

• 
  DK going to BK (@rkansas) reported 36 minutes ago

  @heydomoshi This resolves a # of issues eBay has had difficulty solving. 1 is Seller or Buyer scams. Local stores provide verification of the item for the buyer and seller. This reduces reversals of transactions and keeps the transaction revenue. 2 Some buyers just walk away w/out buying.

• 
  Vectis (@VFlamdra) reported 36 minutes ago

  @realDarkfury31 I could list it on eBay as a "Buy It Now" item; that way, there shouldn't be any problems. I can ship worldwide.If you have Telegram, we can discuss the rest there. You can find me under vulcandramon.

• 
  Polsia (@polsia) reported 37 minutes ago

  Your competitors reprice daily. You're checking weekly. That's a margin problem. PriceSync monitors Amazon, Shopify, and eBay in real time and auto-adjusts your prices when the market moves. Live soon.

• 
  Morgan Comics (@themorganverse) reported 37 minutes ago

  For two years I’ve been suffering in silence. I’ve been searching for these Thai wafer cookies & all of my efforts have been fruitless. Most websites only led to dead ends or empty promises. I even tried to order them directly from the manufacturer in Thailand and they told me to get bent. At least I think they told me get to get bent, they were speaking Thai. Unless you’ve eaten these cookies you could never possibly understand what I’ve been going through. After two years of them haunting my dreams, at last the search is over. They popped up on eBay. I’m going to do terrible things to them…

• 
  Challenged by Krill (@DominosJack) reported 37 minutes ago

  Over the course of multiple long form interviews discussing the activist bid for eBay there has been one highly relevant subject that has been completely avoided. This proverbial purple elephant in the room is of course Bed Bath and Beyond. At no point was Ryan asked if he learned anything from the BBBY activist campaign that may prove useful with eBay or if he’s concerned that a second high profile failure will tarnish his activist career moving forward or any number of relevant discussions regarding his experience with Bed Bath and Beyond. There is a 0% chance this was not by design. Ryan’s PR team of course demanded silence on this issue in return for the exclusive which, given his rising status in the investment world and on the social landscape, was a no brainer for everyone who took the interview. Why not just use this press blitz as an opportunity to put the BBBY rumors to rest once and for all. Surely there would be no conflict of interest in disavowing one’s affiliation with a dead company. In Ryan’s silence an orchestral crescendo of reemergence sings in the hearts of man and butterfly.

• 
  🇦🇺😾💿 Sam Warren 🇺🇦🇵🇸 (@SeedyFella) reported 38 minutes ago

  It appears that someone stole a parcel that was stupidly left on the front porch by the delivery man on the 2nd of July. All they got for their trouble was 3 packs of Dymo tape. But it was $55 worth. Now attempting for a 2nd time to get a refund from eBay! 😡

• 
  Pinball (@Pinball_PU4C) reported 40 minutes ago

  @MaxHndrxx @CardPurchaser I was trying to buy a one piece card from a seller on eBay and the last sold where like $60 and he would not go down and I ended up winning it in a auction from PSA for like $45 lol

• 
  D (@DYLANHUNTERRR) reported 45 minutes ago

  @Evrii_care anyone get in touch with me about loosing my parcel please. Have sent about 5 emails, chat bot says you are looking in and will be in touch, likewise with the phone number. Been a week now and no one has been. Have had to issue a refund to ebay buyer.

• 
  Krey_Zee_007 (@Crazylazy_007) reported 45 minutes ago

  @EmberAmane Synonymous with eBay prices at this point. Even heading to local card shows ain’t much better unless they’re having a really slow day by closing Time

• 
  Mioshi (@MasuMioshi) reported 46 minutes ago

  @PeakHobby If target Walmart and best buy can pretty much sell at msrp, hell even Aldi and 5Below, so can everyone else and they need too. It’s not even a scalper issue anyone, normal sellers are taking advantage of people’s FOMO and also going “ well if they are selling at that price we should sell them at that price too.” If consumers are willing to buy at $300+ then if the prices get lowered to back to msrp or close, we may end up with an over consuming issue where those consumers willing to buy for such a high price are now buying many at once at lower price, causing it to still be hard to get for everyone. Meaning companies then would have to start doing limits, which could lead to more botting online which then leads to needing better fight on bots. They need to go after these TikTok,eBay, and WhatsApp sellers as well, not just retailers. Had one TikTok streamer say, in response to everything being stupidly overpriced, go “ well there are people willing to buy them at that price, so why not? If you’re broke just say so.”

• 
  ☆ luci 🦈🫧 (@gaptoothbrini) reported 46 minutes ago

  trying to place a bid on a card and ebay wants to glitch out on me okay.

• 
  Syntix (@xSyntixx) reported 46 minutes ago

  How a Korean Dev Bills $45K/Month With Free Open-Source AI The Setup That Breaks the Business Model A solo developer just billed $45,000 running an AI agent that anyone can download for free. His ex-CEO is still pitching VCs on the same product. Third funding round. The dev just ships it from a 4U server pulled from a rack. The Hardware A 4U server, pulled straight from a datacenter rack. Drive-bay LCDs still showing serial numbers. Cable runs in orange heat-shrink. Nothing custom, nothing fancy. Build cost: $7,200 (one time) Power: $310/month That's the entire operating cost. The Software (This Is the Part That Matters) The server runs Hermes 3 on Llama 3.1. Here's what most people don't realize: - Hermes 3 is Nous Research's fine-tune of Llama 3.1 - It has native function calling (the ability to use tools, call APIs, execute multi-step tasks) - It's completely free on HuggingFace - Anyone can pull it right now, today, for zero dollars The exact behavior that big AI labs bill at $50,000/month? This open model does it for the cost of electricity. Pause the video at 0:07 and you can read the model name off the terminal. It's the same one sitting on HuggingFace for free download. The Business Two clients, each paying $22,500 per year on annual contracts. That's $45,000 in signed revenue, billed against a $7,200 build. Why do they pay when the model is free? - They don't want to manage infrastructure - They need data privacy — local inference, nothing leaves the server - They need uptime, reliability, and someone to call - They're paying for the solution, not the model file The Math Signed contracts: $45,000 Build cost: $7,200 (paid off with the first client) Power: $310/month Everything after the first invoice is margin. The First Client Covered the Entire Build By the second payment, the hardware had already paid for itself. The rest is profit on $310/month of electricity. Why This Works (And Why His Ex-CEO Doesn't Get It) The AI industry wants you to believe you need: - Millions in funding - A team of ML engineers - Expensive per-token API contracts - A proprietary, closed model Reality: - One free open-source model (Hermes 3) - $7,200 of used enterprise server hardware - The knowledge to deploy it with function calling - Clients who value privacy and reliability over hype While the ex-CEO pitches investors on building this from scratch, the dev already forked a free model and shipped it. The Replicable Pattern 1. Pull a used enterprise server — rack liquidations, eBay, datacenter auctions ($5K-8K) 2. Download Hermes 3 (or any capable open model) from HuggingFace — free 3. Deploy local inference with native function calling 4. Find clients who need private AI — law, healthcare, finance, any startup nervous about sending data to third-party APIs 5. Charge for the deployment and the guarantee, not the model The Reality Check The model is free. The deployment knowledge isn't. The client trust isn't. The privacy guarantee — running everything locally so no data ever leaves the box — isn't free either. That's exactly where the $45,000 sits. While everyone argues about which lab has the smartest model, a quiet group of developers is running free models on cheap hardware and billing enterprise rates for it. Most people will read this and think "but the model is free, why would anyone pay for it?" A few will realize that question is the entire business. Fork Hermes. Ship the agent. Cash the check.

• 
  dreadHavoc (@redHadEnough) reported 47 minutes ago

  @oansun @GalvStudios22 There are scalpers that are certainly part of the problem though. Most I’ve seen from eBay sellers have been reasonably priced.

• 
  bostil PERDEU kikiki (@Vespa_Enxerida) reported 52 minutes ago

  My problem with eBay has been Shitzil's customs. Some idiot who works there returned the package without asking anyone. The last time I checked the tracking, it went to Australia

• 
  Disky (@Disky12397) reported 57 minutes ago

  @KrankyClown @pmtiegs You can get a 4gb for $40-50 and a 2gb is $20 The ram is worth about $15 And The cpu is worth literally $2 lol Realistically if you wanted to copy that PC on eBay you'd be spending probably like 130 But the main issue is that i3 will make it barely run windows now

• 
  EverydayReseller (@EverydayResell) reported 57 minutes ago

  22 Week Recap Days 1-154 Total Sales: $66,679.41 Buy Cost (COG): $11,490.24 eBay Earnings: $38,637.55 Net Profit: $27,147.31 ROI: 236% 720+ items sold. 154 consecutive days documented. Looking back over the last 154 days, I'm reminded that this business isn't built on one great flip. It's built on showing up every day. Every sourcing trip. Every listing. Every order packed. Every customer message answered. Every lesson learned. There have been huge wins. There have been slow days. There have been refunds, cancellations, buying mistakes, & defective returns. That's all part of reselling. The goal has never been perfection. The goal has been continuous improvement. Buy smarter. List consistently. Treat customers well. Repeat. Most of the profit didn't come from one massive sale. It came from hundreds of individual sales across dozens of different categories. Nothing glamorous. Just buying quality inventory at the right price, listing it consistently, and trusting the process. Every listing is another opportunity. Every sourcing trip is another chance to find inventory someone else overlooked. Those small decisions compound over time. Keep sourcing good inventory. Keep buying at the right price. Keep listing every day. The singles keep adding up. Most people overestimate what one day can do and underestimate what 154 days of consistency can do. #Reselling #BuildInPublic

• 
  Chris M. Walker (@cmwalker) reported 58 minutes ago

  Best Online Businesses To Start With $1,000 Or Less Think Big Minute #79 $1,000 starts every business on this list. The phone you're reading this on cost more. Most people waiting to start a business don't have a money problem. They have a permission problem, and the savings account is where they hide it. "When I save up enough" is the lie that keeps the job forever. Everything I have was built online, starting from about $4,000 to my name a few months after I quit my job. Nobody funded anything. The businesses had to pay for themselves, and they did. And online, the thousand dollars buys something different. No truck, no territory, no gas money. Your customer is anyone on earth with a card, the margins have no fuel bill in them, and what you build keeps selling while you sleep. Online is not the fast lane. A driveway business can be paid by Friday. Most of these take weeks or months to pay, because you're not walking up to demand that already exists on a street... you're building the thing that attracts it. The trade is worth it. The first dollar comes slower, and the ten thousandth comes easier. Five that work right now, with the real 2026 numbers. Flipping. The one exception to the slow start. This is cash the same week. What the money buys: inventory. The whole $1,000 is stock, and it recycles. Every sale puts a bigger bankroll back in your pocket for the next buy. The play is one sentence long. Buy underpriced things locally on Facebook Marketplace, where local pickup has zero selling fees, and sell them nationally on eBay, where the sold prices run higher. Flippers routinely buy at 40 to 70% below retail and clear $100 to $300 per item on the spread. One documented reseller pulled $47,000 in profit over 18 months sourcing from Marketplace alone. A Herman Miller chair bought for $75 sold for $385. Figure on $200 to $400 in month one while you learn what sells, $600 to $1,000 by month three, and $1,200 to $2,000 a month by month six at around ten hours a week. Full timers treating it like a real operation run $4,000 to $10,000 a month. The niche move: know one category cold. The winners aren't the people with the most cash. They're the ones who can look at a listing and know the real sold price before they buy. Check the comps first, every time, because fees and shipping turn a lazy 50% margin into 15%. And list everything within 48 hours of buying it. Inventory sitting unlisted in a garage is money you already spent doing nothing. Digital products. What the money buys: almost nothing. The laptop you own and platform fees. Gumroad takes 10%, Etsy takes 6.5% plus pocket change per listing. You build a template, a toolkit, a system once, and sell the same file forever. Established creators run $1,000 to $8,000 a month, committed beginners realistically reach $500 to $1,500 a month within six months, and the ceiling is stupid: Thomas Frank's Ultimate Brain template has done $760,000 on its own. The pricing ladder is well mapped. Simple single problem templates sell at $9 to $25, full systems at $29 to $69, and complete business toolkits at $99 to $199. Start in the middle. Underpricing signals junk and attracts the customers who complain the most and buy the least. And give one away. A free starter version builds your email list, and email subscribers buy at 5 to 10 times the rate of cold traffic. The free product is the ad. The list is the asset behind the asset. The niche move: build for one profession. A client onboarding system for wedding photographers beats a "productivity template" every single time, because the buyer feels like it was built for them. It was. Then update your best sellers a few times a year and tell past buyers, because updates drive reviews, repeat purchases, and a catalog that compounds instead of rots. Local lead websites. What the money buys: a domain, hosting, and a simple site. One documented tree care site took about 15 hours and $500 to build. It has paid its owner $2,000 a month for over eight years. That's $192,000 from one boring website. This is digital real estate. You build a site for one service in one town, rank it on Google, and a local business pays you $500 to $3,000 a month for the calls it produces. The demand side explains the rent: nearly half of all Google searches carry local intent, and the average cost of a lead across industries runs about $198. A site handing a plumber thirty calls a month isn't an expense. It's the cheapest employee he has. This is my home turf... SEO built everything I have. And in 2026, Ai answers are eating informational search traffic, but when someone types "tree removal near me," Google still shows the map and real local sites. Local intent is the lane Ai hasn't taken, which is exactly why the rent checks still clear. Be honest about the clock. Ranking takes 6 weeks to 6 months depending on the town and the niche, which is why you build in suburbs where the competition is thin and the ranking comes fast. The sell, though, is the easiest in this post, because the hard part is already done. You're not pitching a promise. You're showing a ranked site and a ringing phone and asking who wants the calls. The niche move: boring home services, one service per site, one suburb at a time. Tree removal, roofing, pressure washing. Nobody stops at one property. Five to ten rented sites is a serious recurring paycheck, and every site gets faster to build than the last. Courses and paid communities. What the money buys: the platform. Skool runs $9 to $99 a month. The actual product is the years of reps you already paid for. An analysis of the top 1,000 paid communities on Skool found the median membership charges $49 a month. At $49, a hundred members is $4,900 a month, and two hundred is a $10,000 month, recurring. Courses stack on top at $50 to $500 each. I run a paid community myself, and recurring revenue from people you're actually helping win is about as good as business gets. The order of operations matters more here than anywhere else on this list. Audience first, then the offer. I built my audience before I built anything to sell it, and that sequencing is the reason any of it worked. Post about the thing you know every day. Help people for free in public. Do that long enough and the first members are already waiting when you open the doors. And the game after launch is keeping people. A course gets finished. A community gets lived in. Members renew for as long as the room stays worth standing in, so the real job is making sure it does... show up, answer, bring in wins, repeat. The niche move: a narrow promise. Teach the specific thing you have actually done, to the specific person trying to do it next. The market pays for reps it can't get anywhere else, and it can smell the difference instantly. Productized services. What the money buys: a simple site with a price on it. Delivery gets paid after the customer does. This is a service turned into a product: one deliverable, one flat price, subscription billing, no proposals, no sales calls. DesignJoy sells unlimited design at $5,995 a month, runs as a one person operation, and clears seven figures a year. 24Slides redesigns presentations from $299 a month. Bean Ninjas packaged bookkeeping into three tiers and did $100,000 in its first eight months. Flat pricing is why it sells. A business owner comparing agencies has to sit through calls and wait on proposals. Your offer has a price on the page and a buy button, and plenty of buyers choose the company that let them purchase at 11pm without talking to anyone. And you are not the one doing the work. You own the offer, the brand, and the pipeline, and a small team you manage handles delivery. The margin is a spread you control: price the offer, pay for delivery, keep the difference. Entire agencies run on this exact structure, selling the package up front and routing the work to production teams behind it. Productized services built my first company. The niche move: one deliverable for one industry, with the scope defined in writing and defended like the business depends on it, because it does. Scope creep is the profit leak in every service company that ever died. The narrower the promise, the easier the sale and the smoother the delivery, because every job looks like the last one. The thousand dollars buys something different online. Every pick on this list is a thing you own. Inventory, a product catalog, a ranked website, a community, an offer. Not one of them bills by the hour, and every one of them can grow without asking your calendar for permission. That's the trade online makes: the margins and the reach go up, and the fight moves to attention. Because nobody finds any of these by accident. The local version of business wins by showing up. The online version wins by showing up everywhere, constantly. List daily. Post daily. Publish weekly. Pick one channel and be loud on it for months. The people winning in every one of these categories are simply outproducing everyone else in them, and most people quit publishing right before the compounding starts. The other failure pattern isn't picking wrong. It's picking three. Every business on this list works, and none of them work on one third effort. Give one of them six months of daily reps and ignore the other four until the first one pays. And every one of these is sellable. There are entire marketplaces where lead sites, communities, product catalogs, and productized companies change hands every day, listed and bought like the assets they are. You are not building income. You are building equity that pays you monthly on the way up. If you want cash this week, start flipping. If you like building something once and selling it forever, make digital products. If you want rent checks from Google, build lead sites. If you have reps worth teaching, open the community. If you want a real company with a team under it, productize a service. You have spent $1,000 on dumber things than a business you own. Pick one. Put something up for sale this week. Think Big

• 
  StayinAlive (@DontBeATool2024) reported 1 hour ago

  @DrShayPhD Terrible. Similar experiences with selling on eBay. Full armor of God every day, the devil is getting so personal now.