AWS vs Google Cloud SLA

Promised uptime per service, credits, support and what the status pages showed, side by side.

Change vendors (up to four; 2 picked)

VendorHighest SLAWith an SLAFastest support

Promise vs reality, last year

ServiceDowntime vs allowed

Promised in the SLAObserved on the vendor's status pagePartial historyShortfall

Category	AWS	Google Cloud
Compute	99.99%Amazon EC2	99.99%Compute Engine
AI models	99.9%Amazon Bedrock	99.9%Vertex AI Platform (training, prediction, pipelines)
Databases	99.999%Amazon DynamoDB	99.99%Cloud SQL
Storage	99.9%Amazon S3 (Standard)	99.95%Cloud Storage
Data platforms	Not on file	99.99%BigQuery

Vendor and planFastestTargets

AWS Business Support+30 minBusiness-critical system down: < 30 mins · Production system down: < 1 hour · Production system impaired: < 4 hours · System impaired: < 12 hours · General guidance: < 24 hours
AWS Enterprise Support15 minBusiness/mission-critical system down: < 15 mins · Production system down: < 1 hour · Production system impaired: < 4 hours · System impaired: < 12 hours · General guidance: < 24 hours
AWS Unified Operations5 minMission-critical system down: < 5 mins from Incident Management Engineer · Production system down: < 1 hour · Production system impaired: < 4 hours · System impaired: < 12 hours · General guidance: < 24 hours
AWS Developer (legacy, closed to new subscriptions)n/aNo published response targets.
AWS Business (legacy, closed to new subscriptions)1 hCritical issues: 1 hour
AWS Enterprise On-Ramp (legacy, being deprecated)30 minCritical issues: 30 minutes
Google Cloud Standard4 hP2: 4 hours · P3: 8 hours · P4: 8 hours
Google Cloud Enhanced1 hP1: 1 hour · P2: 4 hours · P3: 8 hours · P4: 8 hours
Google Cloud Premium15 minP1: 15 minutes · P2: 2 hours · P3: 4 hours · P4: 8 hours