SLAs, limits, retention periods, IOPS and exam weights.
Every number here already appears on the page named in its Source column. If a figure on this page disagrees with that page, the topic page wins and this page needs fixing.
| Number | Meaning | Source |
|---|
| 100 minutes | Exam time for the question set; reserve about 10 minutes for review | exam-technique.html |
| 700 | Scaled passing score, not 70% | exam-technique.html |
| 30–35% | Infrastructure domain weight | infrastructure/index.html |
| 25–30% | Identity, Governance & Monitoring domain weight | identity/index.html |
| 20–25% | Data Storage domain weight | data-storage/index.html |
| 15–20% | Business Continuity domain weight | business-continuity/index.html |
| Number | Meaning | Source |
|---|
| 99.99% | Availability zones; Standard Load Balancer | cram-sheet.html; infrastructure/load-balancing.html |
| 99.95% | Availability set; AKS with availability zones | cram-sheet.html; infrastructure/containers.html |
| 99.9% | Single Premium SSD VM; AKS without availability zones | cram-sheet.html; infrastructure/containers.html |
| 95% | Single VM on Standard HDD OS disk | business-continuity/index.html |
| < 10 ms | Cosmos DB point reads and writes (SLA) | data-storage/cosmos-capacity.html |
| Spot: no SLA | Spot VMs have no SLA and can be preempted; Azure tries to restore only during the first 48 hours | infrastructure/batch.html |
| Number | Meaning | Source |
|---|
| 1–35 days | SQL point-in-time restore range | business-continuity/database-bcdr.html; cram-sheet.html |
| Up to 10 years | SQL long-term retention | cram-sheet.html |
| 7 / 30 / 35 days | Cosmos DB continuous backup (PITR) windows | cram-sheet.html |
| Every 4 hours | VM backup frequency under Enhanced policy | cram-sheet.html |
| 14 days default, up to 180 | Soft delete retention for backup vaults | business-continuity/backup-vaults.html |
| 360 days / 10 years | Blob operational backup (restores to source) / vaulted backup (restores to another account) | cram-sheet.html |
| 1–365 days | Blob soft delete retention range | data-storage/blob-protection.html |
| 7–90 days (default 90) | Key Vault soft-delete retention, fixed at creation | identity/key-vault.html |
| Every six months | At least one test failover | cram-sheet.html |
| Up to 100 instances | Recovery plan size | cram-sheet.html |
| Number | Meaning | Source |
|---|
| 30 / 90 / 180 days | Minimum retention: Cool / Cold / Archive | data-storage/blob-tiers.html; cram-sheet.html |
| Up to 15 hours | Standard rehydration from Archive (objects under 10 GB); high priority under 1 hour | data-storage/blob-tiers.html; cram-sheet.html |
| 10 GiB/h per account | Archive rehydration throughput | cram-sheet.html |
| 400,000 IOPS | Ultra disk maximum | cram-sheet.html |
| 80,000 IOPS | Premium SSD v2 maximum | cram-sheet.html |
| 65,536 GiB | Maximum managed disk size | data-storage/managed-disks.html |
| 7 days, 90 days | Event Hubs retention range (cram: 1 / 7 / 90 days) | infrastructure/messaging.html; cram-sheet.html |
| 100 MB | Largest Service Bus Premium message | decision-guides.html |
| 30 seconds | Spot VM eviction notice | flashcards.html |
| Number | Meaning | Source |
|---|
| 4 TB | Above this, use Hyperscale | cram-sheet.html; decision-guides.html |
| 15 minutes | Minimum auto-pause delay for serverless (General Purpose, GA) | cram-sheet.html; data-storage/sql-tiers.html |
| 10 databases; 100,000 vCore seconds; 32 GB data and 32 GB backup each | Free SQL Database offer per subscription, per month | cram-sheet.html; data-storage/sql-tiers.html |
| 500 databases, 32 TB | SQL Managed Instance limits | data-storage/sql-platform.html |
| 20 GB | Logical partition size before a hierarchical key is needed | data-storage/cosmos-consistency.html |
| About 50 GB | Physical partition size before a split | data-storage/cosmos-consistency.html |
| 65,536 GiB | Managed disk maximum size | data-storage/managed-disks.html |
| Number | Meaning | Source |
|---|
| 30 instances | Above this, App Service Environment v3 (up to 200 across plans) | cram-sheet.html; infrastructure/app-service-functions.html |
| 1,000 instances | Event-driven scale for Functions | infrastructure/app-service-functions.html |
| 1,000 nodes | AKS dev/test cluster limit (no financially backed SLA) | infrastructure/containers.html |
| 100 Gbps / 400 Gbps | ExpressRoute Direct dual port sizes (10, 100, 400 Gbps) | infrastructure/hybrid-connectivity.html |
| 5 | Maximum diagnostic settings per resource | cram-sheet.html |
| 100 Gbps / 50 Gbps | NAT Gateway StandardV2 / Standard throughput | cram-sheet.html |
| 1,000 / 3,000 | Virtual Network Manager hub-and-spoke spokes / mesh group VNets | flashcards.html |
| 250 Mbps | Azure Firewall Basic throughput ceiling (small business) | flashcards.html |
| 10 Gbps | Azure Firewall Premium single-flow ceiling (fat flow) | flashcards.html |
| 5 seconds | Cross-region load balancer health probe interval (about) | infrastructure/load-balancing.html |
| 2 minutes | Hybrid identity sync interval (about) | identity/hybrid-identity.html |
| Number | Meaning | Source |
|---|
| 1–24 hours | PIM activation duration range | identity/pim.html |
| 14 days | Window to register for MFA (bypass allowed meanwhile if MFA isn't required) | identity/conditional-access.html |
| 30 days (some tables 90) | Default Log Analytics retention | identity/log-analytics.html |
| Up to 2 years | Analytics retention (full KQL) | identity/log-analytics.html |
| Up to 12 years | Total Log Analytics retention | cram-sheet.html; identity/log-analytics.html |
| 7 days or more | Minimum Event Hubs retention for SIEM streaming | identity/log-routing.html |
| 90 days | Activity log kept in the portal; query older history from a Log Analytics workspace | identity/log-routing.html |