AData
API
A defined way for one program to ask another for data or to make it do something.
In practiceOur automation asks the Notion API to create a page whenever a form arrives.
Skip to content
Resources
Platform limits and practical guides.
The numbers and behaviors that decide whether an automation holds up, collected from each platform’s published documentation. Kept current as platforms change.
Cited from official docsKept currentFree
Always confirm before relying on a number
Platforms change limits and plans. These figures reflect each platform’s documentation as we read them. Check the platform’s own documentation before you design around a specific value.
Limits at a glance
What each platform allows.
Three platforms that most automations touch, with the limits that matter and what each means for your design.
| Topic | What to know |
|---|---|
| Request rate | About 100 requests per second in live mode and 25 per second in sandboxes, for most endpoints. |
| Webhook delivery | Failed deliveries are retried for up to three days in live mode. Events can arrive more than once and out of order. |
| Event history | The events list covers the last 30 days. Anything older has to be stored by you. |
| Idempotency keys | Up to 255 characters, and Stripe may remove saved keys after at least 24 hours. |
Design implication: treat every webhook as possibly duplicated or late, check the object’s current state, and use idempotency keys for writes. Read more in Stripe webhooks: retries, duplicates and order.
| Topic | What to know |
|---|---|
| Request rate | An average of about three requests per second per connection, with bursts allowed. Exceeding it returns a rate-limited response. |
| Text size | A single rich-text object holds up to 2,000 characters. |
| Array size | Arrays in a request are limited to 100 elements. |
| Request size | At most 1,000 blocks and about 500 KB per request. |
| Webhooks | Signed with a header you must verify, and retried on failure. |
Design implication: chunk long text, paginate big lists, queue requests and verify webhook signatures. See Notion API limits that break automations.
| Topic | What to know |
|---|---|
| Sheets API quota | About 300 read and 300 write requests per minute per project, and 60 per minute per user. |
| Apps Script runtime | Six minutes per run; custom functions stop at 30 seconds. |
| Daily trigger runtime | About 90 minutes on consumer accounts and 6 hours on Workspace accounts. |
| Email and fetch | Roughly 100 emails a day on consumer accounts versus 1,500 on Workspace, and 20,000 versus 100,000 URL fetches. |
Design implication: batch writes, avoid per-row calls and design triggers around quotas. See Google Sheets automation limits.
Reading a limit
Five questions to ask about any limit.
A number on its own is not useful. These five questions turn it into a design decision.
- 1
Per what?
Per second, minute or day? Per user, project or connection? The unit changes everything.
Unit - 2
What happens at the limit?
Rejected, delayed, queued or silently dropped? Each needs a different design.
Behavior - 3
Does my plan change it?
Higher plans often raise limits. Confirm yours.
Plan - 4
Can it be raised?
Some platforms raise limits on request. Ask before redesigning.
Option - 5
What is my peak?
Limits bite on your busiest day, not your average one.
Peak
Glossary
Terms in plain English.
20 words you will meet in quotes, handover documents and platform docs, each with a one-line definition and an example from a real automation. Search, or filter by topic.
BReliability
Backoff
Waiting progressively longer between retries so a struggling system has time to recover.
In practiceA failed call is retried after 1, 2, 4 and 8 seconds instead of ten times in a row.
CAccess
Customer dashboard
Your private area on BoldBellows where you send logins and API keys, share files, message the team and follow milestones.
In practiceRequests for access appear there, and you answer them there, never by email or form.
Read moreEData
Endpoint
A web address that accepts requests. A webhook is delivered to an endpoint, and an API is called at one.
In practiceThe address we give you to paste into a platform’s webhook settings is an endpoint.
FData
Field mapping
The agreed table that says which field in one system corresponds to which field in another.
In practice‘Customer email’ in the store maps to ‘Contact email’ in the CRM, and ‘Total’ maps to ‘Deal amount’.
Read moreHBuilding
Handover document
The written record you receive at delivery: what was built, every connection and key used, and how to revoke each.
In practiceIt tells you exactly which keys to rotate on the day the project ends.
Read moreIReliability
Idempotency
Designing an action so that doing it twice has the same result as doing it once. It prevents duplicates when events are retried.
In practiceIf the same payment event arrives twice, only one row is created because the event id is checked first.
Read moreLAccess
Least privilege
Giving an account only the permissions it needs to do its job, and nothing more.
In practiceThe automation reads orders and writes to one sheet, so it gets no access to the rest of the drive.
Read moreMBuilding
Milestone
A defined stage of your project with its own due time. Each one has a countdown timer in your dashboard.
In practice‘Build and test’ is milestone two of four, and its timer shows how long is left.
Read moreOAccess
OAuth
A standard way to let an app access your account with limited permissions and no password changing hands.
In practiceYou approve a consent screen listing what the app can read, and you can remove it later from the platform.
Read morePData
Pagination
Reading a long list one page at a time, following a cursor until there are no more pages.
In practiceA store with 8,000 products is read 100 at a time, 80 pages in all.
PData
Payload
The data carried inside a message, such as the body of a webhook, a request or a response.
In practiceA payment webhook’s payload includes the amount, the currency and the customer reference.
PData
Polling
Checking a system on a schedule to see whether anything changed, instead of waiting to be told by a webhook.
In practiceA job that asks a sheet for new rows every five minutes is polling.
QReliability
Queue
A waiting line for jobs, so bursts are processed at a safe pace instead of all at once.
In practiceA flash sale sends 500 orders in a minute; the queue feeds them to the CRM at the rate it can accept.
RReliability
Rate limit
The maximum number of requests a platform lets you make in a period. Go over it and requests are rejected or delayed.
In practiceNotion allows an average of about three requests a second, so a bulk import is paced instead of fired at once.
Read moreRReliability
Reconciliation
Comparing two systems to confirm they agree, then finding and fixing any difference.
In practiceEach night a check compares paid orders with sheet rows and flags any that are missing.
RAccess
Restricted key
An API key limited to specific permissions, so that a leak cannot do everything.
In practiceA key that can read payments but never create refunds, entered in your customer dashboard.
Read moreRBuilding
Revision round
One batch of changes to something we built, sent together and applied together. Two rounds are included.
In practiceYou collect five tweaks, send them as one request, and we apply and test them at once.
Read moreSBuilding
Sandbox
A safe copy of a platform for testing, where mistakes cost nothing and touch no real customers.
In practiceWe run every route against Stripe’s test mode before it touches a live payment.
WData
Webhook
A message a platform sends to a web address of yours the moment something happens, such as a payment succeeding.
In practiceStripe posts to your endpoint when an order is paid, and the automation adds the order to your sheet within seconds.
Read moreNo term matches. Try a shorter word, or ask us on the contact form.
Guides
Read next.
Five long-form articles, each built around a problem we see in real automations.
ArticleStripe webhooks: retries, duplicates and orderWhy a perfect webhook handler does not exist, and what to build instead.Open this pageArticleNotion API limits that break automationsRate, size and block limits to plan for.Open this pageArticleGoogle Sheets automation limitsQuotas, script runtimes and trigger gotchas.Open this pageArticleHow to write an automation briefThe details that decide price and delivery.Open this pageArticleWhat to automate firstA simple way to pick your first workflow.Open this pageArticleAll articlesBrowse every guide.Open this page
Tools
Small helpers.
Free, in-browser, no sign-up. Open any of them from here.
TemplateProject brief templateCopy and fill in.Open this pageDeliveryDelivery estimatorCount business days to a date.Open this pagePolicyRefund checkerSee where your order stands.Open this pagePolicyRevision sorterIs my change included?Open this pageAccessAccess guideWhat each platform needs.Open this pageSearchSearch the catalogFind a service by platform or word.Open this pageMapSitemapEvery page and platform in one list.Open this pageAnswersFAQ searchMore than a hundred answers.Open this pageSecureClient hub chooserFind the right next step.Open this page
Resources FAQ
Questions about our guides.
How we gather, date and update the facts on this page.
Where do your platform facts come from?
From each platform’s published documentation, read and checked by the team, with the source named so you can verify it. We do not rely on forum posts or memory for figures. Where documentation is unclear, we say so instead of guessing.
How often do you update them?
When a platform changes something we have documented, and on a regular review of every figure against the official source. If you notice something that looks out of date, tell us through the contact form and we will check it.
Can I rely on the numbers?
Use them as a guide, then confirm in the platform’s documentation for your plan. Limits depend on plan, region and date, and platforms change them. The only binding figures are those in your accepted quote or order.
Why only three platforms here?
We start with the ones most automations touch: payments, databases and spreadsheets. Every service page lists the limits that apply to its platform pair, and we add more platform guides over time. Tell us which you would like next.
Can you add my platform?
Ask through the contact form with the platform name and the limits that matter to you. If many people ask for the same one, it moves up our list, and we can often answer a specific question directly in the meantime.
Are the guides free to share?
Yes, with a link back. Short extracts with a clear credit are welcome, and the copyright policy explains what to ask first, such as republishing a whole guide or translating it.
Do the limits change by region?
Some do. Check the platform’s documentation for your account’s region and plan. Data residency, payment methods and feature availability can differ by region, and we flag the ones that matter to your project in the quote.
What is a quota versus a rate limit?
A rate limit caps speed, for example requests per second. A quota caps volume, for example requests per day. You can stay under a rate limit and still exhaust a quota, or the reverse, so automations are designed with both in mind.
Want it built for you?
Browse fixed-price services across 128 platforms, or request a custom quote.
Honest limitsFixed prices