Lakebase Postgres SLA
Databricks · Databases · promises up to 99.99% · checked 26h ago
Incidents, last 365 days
The Databricks status page has no public incident history feed.
Commitments and credits
TierUptimeCredit of the monthly bill
- High Availability endpointLakebase endpoints with high availability (secondary compute instances' connection strings excluded)99.99%below 99.99%: 5% · below 99.9%: 10% · below 99%: 25% · below 95%: 50%
- Single Compute endpointLakebase endpoints on a single compute instance99.9%below 99.9%: 10% · below 99%: 25% · below 95%: 50%
More in databases
ServicePromisedFirst creditObserved 365dDowntime
Amazon DynamoDBAWS · Databases · Global Tables SLA99.999%First credit10%Observed 365d99.178%All incidents 99.852%
Managed DatabasesDigitalOcean · Databases · Cluster with Standby Nodes99.95%First credit10%Observed 365d100%All incidents 99.992%
Elastic Cloud Hosted (formerly Elasticsearch Service)Elastic · Databases · Platinum or Enterprise, High Availability deployment99.95%First credit10%Observed 365d99.146%All incidents 99.896%
Cloud SQLGoogle Cloud · Databases · Enterprise Plus edition with HA (or read pool with 2+ nodes)99.99%First credit10%Observed 365d99.927%All incidents 99.488%
Azure SQL DatabaseMicrosoft Azure · Databases · General Purpose / Business Critical / Premium / Hyperscale, Zone Redundant99.995%First credit10%Observed 365dNo feedAll incidents 100%
MongoDB Atlas (database clusters)MongoDB · Databases · Dedicated clusters M10 and above99.995%First credit10%Observed 365d99.989%All incidents 99.215%
Terms, exclusions and sources
- Measured
- Monthly Uptime Percentage per Lakebase Endpoint; a minute is unavailable if all continuous connection attempts via the Covered Connection String fail
- Credit cap
- 50% of the endpoint's monthly compute Fees (highest band)
- Claim window
- By the end of the calendar month following the month in which the SLA Event occurred
- How to claim
- Notify Databricks in writing at help@databricks.com with Verification Information (workspace ID, endpoint ID, request logs)
- Not covered
- Force majeure; Customer noncompliance with the Agreement or Databricks support recommendations; Customer acts or omissions incl. overloading the database, ignoring resource health concerns; Customer equipment, software or other technology; Use with a Beta Service
- Status history
- The Databricks status page has no public incident history feed.
- Weighted uptime
- All incidents, weighted: 100% · Downtime
Less than 99.99% but greater than or equal to 99.9% | N/A | 5% ... Less than 95.0% | 50% | 50%
The page prints no version date. The Single Compute column reads 'N/A' for the 99.9-99.99% band, so the committed level for single compute is 99.9% (inferred from the first credit band). Credits are a % of the endpoint's compute Fees, applied to future Platform Services usage and not counting toward usage commitments.
Sources: each vendor's own SLA page (re-read monthly) and status page feed (read every Friday). A summary, not the contract: the contract you sign governs. Logos via logo.dev; trademarks belong to their owners.