Amazon S3
Keep a bucket of your own up to date with every event, as gzip JSONL files written hourly or daily.
Dashboard → Tools → Amazon S3 Exports. Flip the switch on, fill in the bucket, choose a schedule, and Fictura writes a new set of files after each window. While the switch is off, nothing is written. Missed windows export on their own when it's back on.
What you enter
| Field | Notes |
|---|---|
| Bucket | The bucket name (acme-analytics), not its URL. |
| Region | The bucket's AWS region code, such as us-east-1. |
| Prefix | Optional folder inside the bucket, without a leading slash (fictura/). |
| Access key ID and secret access key | An IAM user's key that can s3:PutObject under that prefix. Nothing more is needed. |
| Schedule | hourly or daily. |
The secret is write-only. It is stored encrypted and never shown again, not even to you; the form only says that credentials are saved. To rotate, enter a new access key ID and its secret together. Leaving both blank when you edit keeps the saved pair.
A minimal IAM policy:
{
"Version": "2012-10-17",
"Statement": [
{
"Effect": "Allow",
"Action": "s3:PutObject",
"Resource": "arn:aws:s3:::acme-analytics/fictura/*"
}
]
}What gets written
- Format. gzip-compressed JSONL: one event per line, the same shape as a row of the Events API.
- Windows. Files are cut by when Fictura received each event, not when it happened, so an event that arrives late still lands in the next run instead of being missed.
- Re-runs. A run that is repeated, or that overlaps the last one, rewrites the same files with the same rows. If your loader can read a file twice, de-duplicate on
uuid.
Run history
Under the form, Run history lists recent runs: when each started, the window it covered, its status, and how many rows and files it wrote. A failed run shows the error the bucket returned, usually a permissions problem with the key.
Newer backends only
On a Fictura backend that predates S3 destinations, the setup says so. For a single file instead of a schedule, use Bulk export.