Skip to content

Service level agreement

Tested with: Generated from catalogue/plans.yaml (sla field)

Monthly availability commitment per plan
PlanMonthly availabilityBackupsSupport
FreeNo SLA1 manual snapshotCommunity
SoloNo SLADaily, 7-day retentionEmail, 1 business day
Team99.9%Daily, 14-day retentionPriority email + chat, 4 business hours
Enterprise99.95%Hourly, 35-day retention · PITR 35 days · cross-region copy · immutable (Object Lock)Dedicated, 1 hour (P1), 24×7
  • Availability is measured per instance per calendar month as the percentage of one-minute intervals in which the instance accepted TLS connections and answered a health query at the gateway. Intervals during a customer- initiated pause, resize or in-place restore, during scheduled maintenance announced 72 hours ahead, or caused by the customer’s allow-list, quota or credentials are excluded.
  • Scheduled maintenance happens in the instance’s maintenance window (default Sunday 02:00 to 04:00 in the region’s local time, configurable). Minor and patch upgrades of the engine run there, always after a backup.
  • Secure placement instances have the Team commitment or the Enterprise contract’s, whichever is higher, with no auto-pause and no spot capacity.
Monthly availabilityCredit (percentage of that instance’s monthly compute charge)
Below the commitment but ≥ 99.0%10%
Below 99.0% but ≥ 95.0%25%
Below 95.0%50%

Request credits within 30 days of the month’s end from Billing → Support; credits are computed from our uptime data, applied as account balance and recorded in the audit log.

databasezy.com/status shows per-region, per-tier component status and incident history; subscribe by email, RSS or webhook. Post-mortems for incidents that breached an SLA are published within five business days.