Skip to main content

Incident · Mid

Ticketmaster/Taylor Swift 2022: 3.5 billion requests, 4x the previous peak

Real incident. Key points: overload control and reservations and concurrency.

12 min read · Incident on 2022, lasted · Source: Ticketmaster · Content reviewed on October 23, 2026

Short answer

Real incident. Key points: overload control and reservations and concurrency. The ticket sale for Taylor Swift's Eras tour used Verified Fan, a pre-registration that Ticketmaster describes as designed to identify real people and keep bots away. More than 3.5 million people registered and about 1.5 million received codes to enter the sale.

01 · CONTEXT

A sale with pre-registration, code, and queue

The ticket sale for Taylor Swift's Eras tour used Verified Fan, a pre-registration that Ticketmaster describes as designed to identify real people and keep bots away. More than 3.5 million people registered and about 1.5 million received codes to enter the sale. Historically, according to the company, about 40% of invitees show up and buy.

The company itself calls the predictability and the confiabilidade hallmarks of Verified Fan, and describes pre-registration as a way to help manage high-demand shows.

  1. The fan logs in. Verification is tied to the account and is checked at that moment.

    Architecture diagram: Access and purchase path with verification
    CLVerified fan
    NWFront door and queue
    CPAccount and code verificationverifies the account
    CPPurchase service
    DBTicket inventory
  2. Only verified people can enter the purchase queue.

    Architecture diagram: Access and purchase path with verification
    CLVerified fan
    NWFront door and queueverified-only queue
    CPAccount and code verification
    CPPurchase service
    DBTicket inventory
  3. To complete the purchase, the buyer must also enter the unique code.

    Architecture diagram: Access and purchase path with verification
    CLVerified fan
    NWFront door and queue
    CPAccount and code verificationvalidates the code
    CPPurchase service
    DBTicket inventory
  4. With the code validated, the purchase service records the sale in inventory (illustrative step).

    Architecture diagram: Access and purchase path with verification
    CLVerified fan
    NWFront door and queue
    CPAccount and code verification
    CPPurchase service
    DBTicket inventorysale recorded

02 · THE PEAK

Registered, invited, and requests: different magnitudes

Ticketmaster attributes the unprecedented traffic to the sum of bot attacks and fans without codes, and reports 3.5 billion total requests to the system, four times the previous peak. Live Nation chairman Greg Maffei said the site was supposed to open to 1.5 million verified fans and that 14 million people, bots included, accessed it.

NumberWhat it measuresWho reported it
3.5 millionPeople pre-registered on Verified FanTicketmaster
1.5 millionPeople who received codes for the saleTicketmaster
14 millionPeople who accessed the site, bots includedGreg Maffei, of Live Nation
3.5 billionTotal requests to the system, four times the previous peakTicketmaster

03 · DEGRADATION

Slowing down to keep the system standing

Those who were not verified did not enter the queue, but the overall volume on the site forced the company to slow down the queues to keep them stable. According to the company, a stadium sale usually takes about an hour; this time some sales were slowed and others postponed, at the cost of longer waits for some fans.

The company estimates that about 15% of interactions on the site had problems, including code validation errors that made fans lose tickets that were already in their cart.

  1. Besides verified fans, bots and visitors without codes generated unprecedented traffic: 3.5 billion total requests, four times the previous peak.

    Architecture diagram: Sales overload: bots, queue, and validation
    CLBots and visitors without a code
    CLVerified fans
    NWEntry pointreceives all traffic
    QWaiting queue
    CPCode validation
    CPPurchase service
    DBTicket inventory
  2. Those who were not verified could not enter the queue.

    Architecture diagram: Sales overload: bots, queue, and validation
    CLBots and visitors without a codedoes not enter the queue
    CLVerified fans
    NWEntry point
    QWaiting queue
    CPCode validation
    CPPurchase service
    DBTicket inventory
  3. Even so, the overall volume on the site forced the queues to be slowed to keep them stable, with longer waits for some fans.

    Architecture diagram: Sales overload: bots, queue, and validation
    CLBots and visitors without a code
    CLVerified fans
    NWEntry point
    QWaiting queueslowed queue
    CPCode validation
    CPPurchase service
    DBTicket inventory
  4. Code validation errors made some fans lose tickets that were already in their cart.

    Architecture diagram: Sales overload: bots, queue, and validation
    CLBots and visitors without a code
    CLVerified fans
    NWEntry point
    QWaiting queue
    CPCode validationvalidation errors
    CPPurchase service
    DBTicket inventory
  5. With demand above the remaining inventory, the next public sale was canceled.

    Architecture diagram: Sales overload: bots, queue, and validation
    CLBots and visitors without a code
    CLVerified fans
    NWEntry point
    QWaiting queue
    CPCode validation
    CPPurchase service
    DBTicket inventoryinsufficient inventory

Even so, more than 2 million tickets were sold on November 15. On the eve of the public sale, Ticketmaster announced its cancellation, citing extraordinary demand on its systems and insufficient remaining inventory.

04 · THE TESTIMONY

What the company told the Senate

Joe Berchtold, president and chief financial officer of Live Nation, testified before the US Senate Judiciary Committee. According to reports, he said bot traffic was three times larger than the company had ever experienced and that, for the first time, bots went after the Verified Fan system. He stated that the bots did not penetrate the systems or acquire tickets, but that the attack forced the company to slow down and even pause sales.

According to CFO Dive, the written testimony said the company could improve by staggering sales over a longer period and setting fans' expectations for obtaining tickets.

05 · LESSONS

What the design needs to anticipate

Admission before inventory: when demand can exceed capacity by an order of magnitude, deciding who gets in before touching inventory protects the critical path. A waiting room, account verification, and a per-buyer code are layers of that decision, and a waiting queue only helps if the exit rate respects what the purchase can handle.

Size for the traffic, not for the guest list: bots and visitors without a code also consume capacity, even though they cannot buy. That is why sizing and load tests must start from the worst plausible traffic, not from the number of invitees, and the escalabilidade at the edge counts as much as the one on inventory.

Steps on the critical path must tolerate retries: if a validation fails under load, the legitimate buyer tries again, and the operation must be idempotente and must not discard the cart. The design of seat reservation with locking and expiry is in the classic ticket-office case; this case covers what comes before it.

Frequently asked questions

What is Ticketmaster/Taylor Swift 2022: 3.5 billion requests, 4x the previous peak?

Real incident. Key points: overload control and reservations and concurrency.

What does the Ticketmaster/Taylor Swift 2022: 3.5 billion requests, 4x the previous peak design cover?

It is organized in these sections: A sale with pre-registration, code, and queue, Registered, invited, and requests: different magnitudes, Slowing down to keep the system standing, What the company told the Senate, and What the design needs to anticipate.

How long is Ticketmaster/Taylor Swift 2022: 3.5 billion requests, 4x the previous peak and what level is it?

It takes about 12 minutes to read and is rated Mid level.

Now design it yourself

Run this case as an interview in the System Design school: you make the decisions, the tutor challenges them and you see how your design scores.