This page outlines the current steps for the Mattermost team for creating a data export file for a Mattermost Cloud administrator. All exports are in JSONL format by default.
- Cloud user contacts Mattermost, requesting a data export. This can be made by contacting the support team via https://customers.mattermost.com/cloud/contact-us, or through the Mattermost Help Center at https://support.mattermost.com/.
- Zendesk ticket automatically created for the request, assigned to a Support team member.
- Support team member confirms which of the following data exports is requested:
- Configuration data in JSONL format, which they can apply to their Self-Hosted deployment.
- Configuration and bulk export data with channels, messages, and more in JSONL format.
- Configuration and bulk export data in JSONL format, plus all assets stored in s3 (e.g. files).
- Everything in the database, with either a raw database dump, or pre-formatted CSV file.
- Support team member opens a Jira ticket for the SRE team to process the data export request with the following information (see MM-38454 for an example):
- Cloud URL, e.g. https://example.cloud.mattermost.com/
- Account owner email, e.g. [email protected]
- Data export format, e.g. JSONL
- Cut-off date for history export, e.g. Sep 9, 2021
- Link to Zendesk ticket, e.g. https://mattermost.zendesk.com/agent/tickets/1234
- SRE team performs the data export per request, with a 5-business day SLA, and stores it in S3.
- SRE team shares the data export S3 link to the Support team.
- Support team forwards the S3 link to the requester.
Long-term we'd like to make this process self-serve.
In the near term, we are aiming to publish documentation about Cloud data imports/exports by October 25.
This will make the data export process for channels, messages and other bulk export data self-serve, with some knowledge about running commands to export the data.
Providing a self-serve import/export capability via the UI is not in the near-term roadmap, but can be considered if we see an increased demand. An early draft of this proposed approach can be found here.