Rate limits control how much request traffic a retailer will accept within a given period. Request throttling occurs when traffic reaches or exceeds those limits, which can slow, challenge, reject, or temporarily restrict requests.
The Shit Bot is a sneaker bot and Nike SNKRS bot that also runs as a Pokémon Center bot and Pokémon bot.
For sneaker botting, this matters because task count, delays, sessions, proxies, and the requested endpoint can all affect how much traffic a setup generates. Sending more requests can improve reaction time up to a point, but requests that keep getting throttled aren’t making the task faster.
What Are Rate Limits in Sneaker Botting?
A rate limit is a rule that controls how frequently a client can send requests to a website, API, or particular endpoint.
A sneaker bot can make requests throughout a release. Monitoring may check for product or stock changes, while later requests can handle authentication, carting, queues, shipping, payment, and order submission. Our guide to the sneaker bot checkout flow breaks down how those stages connect.
There’s no universal rate limit across sneaker sites. A retailer can set different thresholds, time windows, and rules for different parts of its infrastructure.
RFC 6585 also leaves the method of identifying and counting clients up to the server. That means an IP address can matter, but it shouldn’t automatically be assumed to be the only characteristic involved.
How Does Request Throttling Actually Work?
Request throttling controls traffic after a configured request condition has been reached.
Take a simplified limit of 20 requests per minute. If requests were distributed evenly, that works out to one request every three seconds. A client sending one request every second would theoretically attempt 60 requests during the same minute, or three times our hypothetical limit.
| Example request pattern | Requests in 60 seconds | Against a hypothetical 20/min limit |
|---|---|---|
| 1 every 6 seconds | 10 | 10 below |
| 1 every 3 seconds | 20 | At limit |
| 1 every 2 seconds | 30 | 10 above |
| 1 every 1 second | 60 | 40 above |
These are example calculations, not published limits for any sneaker retailer.
Actual systems can be more flexible. Cloudflare, for example, documents rate-limiting configurations that can include a matching expression, requests per period, mitigation duration, and characteristics used to count requests.
So request frequency is only one part of the picture.
Why Does Burst Traffic Matter?
Burst traffic is a large number of requests arriving within a short period instead of being spread evenly over a longer window.
This distinction matters around sneaker drops because tasks don’t necessarily generate perfectly even traffic. A setup can suddenly become more active when a product goes live, tasks are started together, stock changes, or multiple tasks advance to another checkout stage at roughly the same time.
Consider two hypothetical clients that each make 60 requests during one minute. One sends roughly one request every second. The other sends 40 requests during the first 10 seconds and only 20 during the remaining 50.
Both total 60 requests, but their traffic patterns are very different.
That’s why requests per minute alone cannot fully describe how a retailer might treat traffic. The timing and concentration of those requests can matter too.
How Do Task Count and Delay Affect Request Volume?
Task count and delay can multiply the number of request opportunities a setup creates.
If one task performs one request every five seconds, the simple theoretical maximum is 12 requests per minute. Ten tasks following that exact pattern would create 120 request opportunities. Twenty would create 240.
Reduce the interval to two seconds and the theoretical number rises again.
| Simplified setup | Theoretical request opportunities per minute |
|---|---|
| 1 task at 5 seconds | 12 |
| 10 tasks at 5 seconds | 120 |
| 20 tasks at 5 seconds | 240 |
| 20 tasks at 2 seconds | 600 |
The calculation is simply: tasks × (60 ÷ delay in seconds).
These figures aren’t estimates of how many requests TSB or another sneaker bot actually sends. A real module can make different requests at different stages, pause, retry, wait on a queue, or behave in ways that don’t map directly to one request per delay cycle.
The table is useful because it shows why delay can’t be considered separately from concurrency. A five-second delay across one task and the same delay across 20 active tasks represent very different potential traffic levels.
Do All Retailer Endpoints Share the Same Rate Limit?
Different endpoints shouldn’t be assumed to share one universal request budget.
A sneaker checkout involves multiple stages, and those stages can communicate with different endpoints. Product monitoring isn’t necessarily handled by the same infrastructure or rules as login, carting, payment, or order submission.
This changes how we should think about throttling.
A task might be able to monitor normally but encounter restrictions after moving into another stage. Likewise, high traffic against one resource doesn’t prove that every other resource on the site has reached the same threshold.
This is one reason the full request path matters when diagnosing bot behavior. Looking only at the total number of tasks can hide where the traffic is actually going.
The sneaker bot checkout flow is useful here because it separates the major stages instead of treating checkout as one request.
Are Rate Limits Based Only on Your IP Address?
Rate limits don’t have to be based exclusively on IP addresses.
RFC 6585 notes that servers can identify users in different ways, including authentication credentials or cookies. MDN similarly notes that rate-limiting implementations vary and may operate server-wide, against a resource, by IP, or according to authenticated users.
That matters because a sneaker bot session contains more state than its current proxy.
Cookies, authentication data, tokens, and other session information can persist while a task moves through a retailer. Our guide to sneaker bot sessions covers how that state is maintained through the shopping process.
This doesn’t mean every retailer rate-limits every one of those characteristics. The exact implementation is generally controlled server-side and may not be publicly documented.
The practical point is narrower: changing an IP shouldn’t automatically be treated as resetting every possible request limit.
Can Multiple Tasks Share the Same Request Budget?
Multiple tasks can potentially contribute toward the same limit if the retailer groups their requests using a characteristic those tasks share.
The easiest example is an IP-based limit. If several active tasks send traffic through the same proxy address, the retailer sees multiple requests associated with that address rather than separate IP identities for every task.
But IP isn’t the only possible grouping mechanism, so the exact request budget cannot be inferred from proxy count alone.
This is also where aggressive concurrency can become misleading. Twenty tasks may look independent inside the bot interface while still sharing network or session characteristics outside it.
TSB’s guide to why sneaker proxies get banned or blocked covers the related proxy side, including how request volume and traffic patterns can affect an address.
A proxy can distribute traffic differently. It cannot guarantee that every task receives an independent rate limit.
What Happens When a Sneaker Bot Gets Throttled?
A throttled request can receive an explicit rate-limit response, but throttling does not have to look identical on every site.
The clearest standard signal is HTTP 429 Too Many Requests. RFC 6585 defines 429 for cases where a client has sent too many requests within a given amount of time. A response can also include Retry-After, which tells the client how long the server wants it to wait before another attempt.
We already cover that status in detail in our HTTP status codes for sneaker bots guide, alongside 403 and 503 responses.
The distinction matters here because this article is about the traffic condition behind throttling, not HTTP codes themselves.
A 429 is strong evidence that request frequency or another quota has been exceeded. Without an explicit response, however, we should be careful about claiming the exact server-side cause from task behavior alone.
Should a Bot Immediately Retry a Throttled Request?
Immediately retrying every restricted request can generate more traffic against the same resource that just rejected it.
Imagine 20 tasks are restricted at roughly the same time. If all 20 immediately retry and are rejected again, another 20 requests have been added without advancing checkout.
If each task repeats that cycle five times, those original 20 failures can produce 100 additional retry attempts.
That doesn’t prove how any specific retailer handles retries, but the arithmetic shows why uncontrolled retry behavior can be counterproductive.
Where a server supplies Retry-After, the response provides explicit information about when it expects another attempt. The header can contain either a delay in seconds or a date, according to MDN.
This is another reason throttling should be treated as a traffic-management problem rather than simply an error that needs to be retried as quickly as possible.
Does Sending Requests Faster Make a Sneaker Bot Faster?
Sending requests faster only helps when those extra requests remain useful.
Suppose a monitor changes from a 10-second interval to five seconds. Under our simplified one-request-per-cycle model, that increases potential checks from 6 to 12 per minute. At one second, it becomes 60.
The raw monitoring frequency has increased tenfold from the first example to the last.
That doesn’t mean useful performance increased tenfold.
If additional requests start receiving restrictions, challenges, or other unsuccessful responses, the lower delay can create more network activity without producing equivalent progress.
The same distinction applies to proxy performance. As we explain in our proxy latency and location guide, a fast network response is useful only as part of a request that successfully reaches and progresses through the retailer’s systems.
The useful metric isn’t how many requests a setup can send. It’s how effectively those requests move tasks forward.
How Can You Tell If Your Setup Is Being Rate Limited?
Rate limiting is easiest to identify when the server explicitly returns HTTP 429, particularly if the response also includes retry information.
Without that signal, look at the pattern rather than one isolated failed request.
A restriction that appears as concurrency increases, affects several tasks sharing characteristics, repeatedly occurs against a particular endpoint, and temporarily changes after traffic falls can be consistent with throttling. None of those observations individually reveals the retailer’s exact rule.
That distinction matters because sneaker sites use other controls too. Our guide to how sneaker websites detect bots covers other signals that can affect requests, including network reputation, browser characteristics, fingerprints, JavaScript results, and session behavior.
Rate limits and request throttling are ultimately traffic controls. For sneaker botting, understanding them means looking beyond a single delay setting and considering task concurrency, request timing, endpoints, sessions, proxies, and retries together.
A lower delay can create more request opportunities. Whether those requests actually make a setup faster depends on what happens after they reach the retailer.
Slow Your Roll: Why Every Request Has Its Limits
In sneaker botting, going full throttle isn’t always the smartest move. More tasks, shorter delays, and more requests can increase activity, but useful speed still comes down to whether those requests actually make progress.
Rate limits make that tradeoff clear. The goal isn’t simply to generate more traffic. It’s to find a setup where your tasks can keep moving without unnecessary requests working against them.
That’s also where TSB fits into the bigger picture. Years of user success across different drops show what can happen when the right setup meets the right opportunity. If you’re learning how to find the right pace for your own setup, your TSB key can be another piece of the ride as you take on future drops.
