Our company is deeply integrated with the Airtable API. In fact, we help other companies build on Airtable - so we monitor the Airtable API closely.

Up until a month ago, all our API requests to Airtable were GET requests. With the launch of our proxy, we’re now servicing PATCH/POST/DELETE requests as well.

With these new requests, we began to notice some unusual API performance. Requests could intermittently take up to 10 seconds to complete.

So we started an investigation to determine what caused this erratic API performance.

Our first hypothesis was that base size was the culprit. We ran a benchmark across bases of different sizes, and oddly the 10K base was much slower. Granted, each base had different data in it - but the consistent performance on the other 1K base and 49K base indicated that size wasn’t the issue.

So it must be something with the structure of the base that impacts API performance…

Read the full post to learn what elements of your base cause the Airtable API to slow down.