Backup and restore

Backups are always enabled on Tembo Cloud instances.

Find Restore in Instance dashboard > Restore from backup


Tembo performs Postgres backups and Write-Ahead Log (WAL) archiving using Barman, powered by the Cloud Native Postgres Operator (CNPG).

Backups are saved to Amazon S3.

Retention policies

  • The retention policy for Barman is configured to 30 days
  • The retention policy for AWS S3 is set to 40 days
  • AWS S3 object versioning is enabled, and the retention policy for non-current versions is 30 days

WAL archiving

Tembo sets archive_timeout to 5 minutes, ensuring that WAL files are closed and archived to S3 at least every five minutes, providing a deterministic time-based value for your Recovery Point Objective (RPO).


Tembo Cloud supports Point-In-Time Recovery (PITR) to a new Tembo Instance for all running Instances.

To restore a running instance, navigate to Instance dashboard > Restore from backup.

Restore via API

For information on authenticating to the API, please see the Tembo Cloud API authentication guide.

TEMBO_TOKEN=<your token>
TEMBO_ORG=<your organization id>
TEMBO_INST_RESTORE_FROM=<your instance id>
TEMBO_INST_RESTORE_TO=<your new instance name>

# Example, recover to five minutes ago
TEMBO_RECOVERY_TIME=$(date -u +'%Y-%m-%dT%H:%M:%SZ' -d '-5 minutes')

curl -X 'POST' \
  "$TEMBO_ORG/restore" \
  -H "accept: application/json" \
  -H "Authorization: Bearer $TEMBO_TOKEN" \
  -H "Content-Type: application/json" \
  -d @- <<EOF
    "instance_name": "$TEMBO_INST_RESTORE_TO",
    "restore": {
        "instance_id": "$TEMBO_INST_RESTORE_FROM",
        "recovery_target_time": "$TEMBO_RECOVERY_TIME"

If recovery_target_time is omitted, then it will restore to the current time.

For more information, please see the Tembo Cloud Platform API documentation.


chevron right arrow

Increasing compute

Configuration and management