API Rate Limits: Why Your Integrations Fail at Busy Times
Short answer: Most connected tools talk to each other through APIs, and every API caps how many requests a single app can make in a given window. When a sale, a newsletter send or a bulk import pushes your integration past that cap, the other side answers "429 Too Many Requests" and stops listening for a while. The fix is rarely a bigger plan; it's usually an integration that waits and retries properly, sends fewer, larger requests, and reacts to events instead of constantly asking "anything new?".
What a rate limit is, without the jargon
Picture a shop counter with one assistant. Customers can queue, but if a coach party arrives at once, some are asked to come back in a minute. APIs do the same. MDN describes the 429 status as the server telling a client it has sent too many requests in a given amount of time, sometimes with a Retry-After header saying how long to wait.
Each platform sets its own rules:
- Shopify uses a "leaky bucket" for most of its APIs. Each app has a bucket whose size depends on the API and your plan; requests add to it, and it drains at a steady restore rate. Short bursts are fine as long as your average stays under that rate. Shopify also says it may temporarily reduce limits to protect the platform, so integrations must handle throttling gracefully.
- HubSpot publishes a short burst limit per 10 seconds and a daily limit, which vary by subscription tier; for privately distributed apps, the daily allowance is shared across all apps in the same account (as of November 2026). That last detail catches people out: three separate tools pulling from one HubSpot account share one daily allowance.
- Google's APIs set quotas per project and per user. The Drive API documentation, for example, notes a 403 "User rate limit exceeded" response when a quota is exceeded, and a 429 from additional backend checks.
The figures change, so check each provider's current limits page rather than trusting a number in a forum post.
Symptom → cause → fix
| What you notice | Likely cause | Fix |
|---|---|---|
| Orders sync fine most days but go missing during sales or after an email campaign | Burst of orders triggers a burst of API calls; the integration hits 429 and drops them | Queue outgoing calls and retry with backoff instead of firing them all at once |
| A nightly sync "partly" completes: some products update, others don't | Sync loops through thousands of single-item requests and runs into the limit halfway | Use batch or bulk endpoints; resume from the last successful item |
| Stock levels lag by 10–30 minutes | Integration polls every minute for changes and gets throttled | Switch to webhooks, so the platform tells you when something changes |
| Errors appear only on the first of the month or Monday mornings | Scheduled jobs from several tools start at the same minute | Stagger schedules; spread heavy jobs across the night |
| CRM sync breaks after a new marketing tool is connected | New tool shares the same account-wide daily allowance | Audit connected apps; remove ones you no longer use; reduce what each one pulls |
| Integration logs show the same request retried dozens of times in a second | Retry loop with no wait, making throttling worse | Exponential backoff with a cap, and respect Retry-After |
| Everything fails at once, then recovers | Provider temporarily lowered limits or had an incident | Check the provider's status page; ensure failed jobs are retried later, not lost |
The fixes, explained
Wait, then retry, properly
When a request is throttled, the worst response is to hammer it again immediately. Google's Drive guidance recommends truncated exponential backoff: retry after a short wait, double the wait each time, stop at a maximum, and give up gracefully after a set number of attempts. Shopify's limits page recommends that code stops making requests until enough time has passed, with one second as its suggested backoff.
If the response includes Retry-After, use that value. It's the provider telling you exactly when to come back.
Ask for less, less often
Shopify's own advice is to request only the data your app needs and cache data you use often. HubSpot's guidelines say the same: cache settings like properties and owners rather than fetching them on every run, and use batch APIs.
A hypothetical example: a Galway gift shop's stock sync fetches every product's full record every 15 minutes. Changing it to fetch only products updated since the last run, and only the stock fields, can cut the number of calls dramatically without changing what the shop sees.
Let events come to you
Polling means asking "has anything changed?" over and over, which burns your allowance even when nothing has. Webhooks flip that: the platform sends a message to your system when an order is created or stock changes. Most platforms offer them; HubSpot, for instance, mentions webhook actions in workflows as a way to send record data out.
Put a queue in the middle
A queue holds jobs (for example "send order 1042 to the accounting tool") and works through them at a safe pace. If the receiving API is busy, the job waits instead of disappearing. If you use a no-code integration platform, find out what it does with a step that fails repeatedly: is it retried, held for you to replay, or silently skipped?
Watch the numbers before they bite
Many APIs report your remaining allowance in response headers. HubSpot, for example, returns headers such as X-HubSpot-RateLimit-Remaining. A good integration reads them and slows down before it hits zero. At the very least, log every 429 and set an alert if they spike.
Questions to ask whoever built your integration
- What happens to a record when the API says "too many requests"? Is it retried or lost?
- Do we poll, or use webhooks?
- Are we using batch endpoints where the platform offers them?
- Where can I see failed syncs, and who is alerted?
- How many apps share our account's allowance?
If nobody can answer the first question, that's the place to start.
FAQ
Will upgrading my plan fix rate limit errors?
Sometimes limits rise with plan tier, as with Shopify and HubSpot. But an integration that polls wastefully will hit the higher limit too. Fix the pattern first.
Is a 429 error my fault or the platform's?
Usually it means your integration asked too fast. Occasionally the provider lowers limits temporarily, which is why every integration should retry later rather than drop data.
Do no-code tools handle rate limits for me?
Not always. Error handling varies by tool and by step, so read each tool's own documentation on failed runs. Check each failed run in the history rather than assuming it recovered.
How do I know which app is using up the allowance?
Some platforms show API usage per app in their developer settings or logs. Otherwise, disconnect unused apps and review what the remaining ones sync.
Got a sync that falls over every time you run a promotion? Tell us which tools are connected and we'll look at where the calls pile up.