monday.com API and webhooks: Real-world examples and use cases for developers
I'm a backend developer building custom integrations for our company's monday.com instance. I need real-world examples of API calls and webhook implementations. Specifically interested in: 1) How to efficiently bulk-sync data without hitting rate limits, 2) Best practices for webhook security and retry logic, 3) Common API gotchas that aren't in the documentation. We're using Python and Node.js.
3 Answers
Here's what I've learned from building multiple monday.com integrations: 1) Rate limits: The API allows 10 requests per second, but use batching for bulk operations - the 'change_multiple_column_values' mutation can update up to 100 items in one call. Cache your board schemas locally to avoid repeated 'fetch_boards' calls. 2) Webhooks: Always implement idempotency - monday.com retries webhooks up to 3 times with exponential backoff. Store webhook events in a queue (we use Redis) before processing. Verify the webhook signature using the secret from the developer portal. 3) Gotchas: Column IDs change when you rename columns - always use column IDs, not names, in your code. The API returns dates in Unix timestamp format, not ISO. Subitems are separate API entities with their own board_id - you can't query subitems in the same request as parent items. Use GraphQL fragments to reduce payload size.
One undocumented gotcha: the API token has different permission levels. Personal API tokens have full access to everything the user can see, but OAuth tokens respect the specific scopes requested. If your integration suddenly can't see certain boards, check your OAuth scopes. Also, for webhooks, the 'event' payload structure changed in the last API update - make sure your parser handles both v1 and v2 event formats during the transition period.
For Python developers, use the 'monday' SDK but don't rely on it exclusively - it doesn't support all query types. We built a wrapper that handles pagination automatically. For Node.js, the official SDK is better but watch out for the 'items_page' query which has a 100-item limit per page. Implement cursor-based pagination properly or you'll miss items. Also, set up proper error handling for the 429 (rate limit) response - implement a backoff strategy that respects the 'retry-after' header.