# Limitations

> Platform-level limits and limitations affecting all Integration Service connectors in Automation Cloud, including the 8 MB data processing cap and API constraints.

This section describes platform-level limits and other limitations that impact all
Integration Service connectors and/or activities.

## Data processing limits

Integration Service has a platform limit for retrieving responses. The maximum limit
for data processing is 8 MB. This means that, for some API calls, the number of
responses retrieved can exceed the 8 MB limit and throw the following error message:
"Response content too large".

:::note
The 8 MB limit applies to data in JSON format. This
includes cases where files are sent or received in a Base64 format as part of
the JSON data. For files handled outside of JSON, the limit is 1 GB.
:::

### Workaround

You can take the following steps to ensure the data processing limit is not
exceeded:

* Adjust the **Max records** field at activity level to a lower
  number.
* Add filtering to your queries. Providing filter parameters can limit the
  number of responses.

## Activity timeouts

There is a platform-wide timeout of 90 seconds for all activities and triggers based on Integration Service connectors. This means that if a response is not received from the API within 90 seconds, the activity fails with a timeout error.

## Trigger limitations

This section lists the known limitations of Integration Service triggers.

### Auto-disabling triggers

If a trigger encounters 100 consecutive errors caused by issues in third-party
applications during polling for new or updated events, the system automatically
disables the trigger. This mechanism helps maintain system stability.

When one of your triggers is disabled, you will receive an email notification. You
must manually re-enable the trigger after the underlying issue causing the errors
has been fixed.

### Data processing limit

Integration Service enforces an **8 MB data processing limit** per event payload. This applies not only to activities but also to **triggers**, meaning large record payloads may fail to process if they exceed the limit.

## Polling interval behavior

The polling interval defined at the **connection level** for **polling-based triggers** (including *Wait for Event and Resume*) is automatically applied to all triggers associated with that connection. This means that:

* **Individual triggers cannot have their own polling intervals.** All triggers under the same connection will inherit the interval configured on the connection.
* Updating the interval on the connection implicitly updates it for all associated triggers.