---
updatedAt: 2026-07-13T14:00:14.000Z
agentTools:
siteIndex: https://developers.pismo.io/llms.txt
projectIndex: https://developers.pismo.io/pismo-docs/llms.txt
---
# Setting up event batch file delivery
In addition to [real-time notifications](https://developers.pismo.io/pismo-docs/docs/setting-up-real-time-event-delivery), event data can be delivered as files via a configured integration channel to cloud storage — Amazon Web Services (AWS), Google Cloud Platform (GCP) — or downloaded using Secure File Transfer Protocol (SFTP).
## How it works
Event data is batched for **5 minutes** or until it reaches **10 MB** in size, then saved as a new object containing event JSON data in the `/main_stream/` path.
For information security, no S3 bucket is ever exposed publicly. A [Key Management Service (KMS)](https://aws.amazon.com/kms/) key is created for each bucket so that stored objects are encrypted and only decrypted when you download them.
Choose your delivery method below and follow the corresponding tutorial.
**File retention and download responsibility** Pismo retains files while your organization is active. However, we strongly recommend downloading files to your own infrastructure as soon as they are produced — either in real time (triggered by event notifications) or within 24 hours if you use a batch/polling strategy. These files belong to your organization, and keeping a local copy ensures you always have access. Please note that Pismo may charge additional fees for recovery requests of files older than 30 days.
# AWS S3
The following image shows the data flow for event file delivery to an AWS account
#### How it works
1. Event files are generated and saved in a dedicated S3 bucket exclusive to your Org.
2. A listener detects the new file and generates a new bucket object event.
3. If you have real-time event delivery configured, the event is sent as a notification to your [SNS topic](https://docs.aws.amazon.com/sns/latest/dg/sns-create-topic.html). Otherwise, you can poll the bucket periodically.
4. Your AWS account receives the event identifying the bucket, object path, and size.
5. Your routine invokes ([`sts:AssumeRole`](https://docs.aws.amazon.com/STS/latest/APIReference/API_AssumeRole.html)) on Pismo's IAM Role, and then downloads the file via ( [`s3:GetObject`](https://docs.aws.amazon.com/AmazonS3/latest/API/API_GetObject.html)).
#### Pismo creates for you:
* **S3 bucket**— Exclusive to your Org. Example: `arn:aws:s3:::pismo-dataplatform-tn-376a4170-3c93-4676-b294-ec0b4241c7ab`
* **IAM Role** — For file downloading (`s3:GetObject`). Example: `arn:aws:iam::{pismo_aws_account_id}:role/dataplatform-consumer-tn-376a4170-...`
**Your configuration:** Follow the AWS event file configuration tutorial.
# GCP cloud storage
The following image shows the data flow for event file delivery to a GCP account:
**How it works:**
1. Event files are generated and saved in a dedicated S3 bucket.
2. A listener detects the new file and generates a notification.
3. The notification is delivered to your [Pub/Sub](https://cloud.google.com/pubsub) topic.
4. Pismo's event consumer copies the file to your GCP Cloud Storage bucket.
**Your configuration:** Follow the [GCP event file configuration tutorial](https://developers.pismo.io/pismo-docs/docs/gcp-event-file-configuration-tutorial).
## SFTP
If you can't use cloud-native integrations (AWS/GCP), or prefer a simpler approach, SFTP is available. Pismo's SFTP service abstracts the S3 bucket behind an SFTP layer — when you connect, you access the same files.
You can associate up to **8 SSH** keys per organization. Sandbox and production credentials are separate.
**Access endpoints:**
| Environment | Endpoint | Port |
| --------------------- | --------------------------- | ---- |
| Sandbox / Development | sftp-main.data.pismolabs.io | 22 |
| Production | sftp-main.data.pismo.io | 22 |
**Knowing when files are available**: Set up real-time event notifications. Otherwise, periodically connect and check for new files.
**Your configuration:** Follow the SFTP event file configuration tutorial.
## Validate integration
Each provider-specific tutorial includes a **Validate Integration** section with:
* Step-by-step manual validation commands
* Automated diagnostic scripts that check authentication, resources, permissions, and connectivity
Follow the tutorial for your chosen delivery method to access these tools.
**Need Pismo-specific details?** Information such as Pismo's AWS Account ID, single-tenant SFTP endpoints, consumer role ARNs, or external IDs can be obtained from your **Technical Account Manager (TAM)** or the **Implementation Engineer** assigned to your project.