All Cheat SheetsExplore Deep Dive Guide
Explore Deep Dive Guide
Explore Deep Dive Guide
Explore Deep Dive Guide
Explore Deep Dive Guide
Explore Deep Dive Guide
Databases Cheat Sheet
Multi-cloud reference for AWS, Azure, and Google Cloud Databases services
Databases — Multi-Cloud Cheat Sheet
Cloud Arena | cloudarena.com
Amazon Aurora / RDS
AWSKey Commands & CLI Operations
| Command | Description | |
|---|---|---|
| aws rds create-db-cluster --db-cluster-identifier prod-cluster --engine aurora-postgresql --engine-version 15.4 --master-username dbadmin --master-user-password SECRET --database-name appdb | Provision Aurora PostgreSQL cluster | |
| aws rds create-db-instance --db-instance-identifier prod-reader-1 --db-cluster-identifier prod-cluster --engine aurora-postgresql --db-instance-class db.r6g.xlarge | Add low-latency read replica | |
| aws rds failover-db-cluster --db-cluster-identifier prod-cluster | Trigger immediate cluster failover | |
| aws rds modify-db-cluster --db-cluster-identifier prod-cluster --serverless-v2-scaling-configuration MinCapacity=0.5,MaxCapacity=32.0 | Enable Aurora Serverless v2 auto-scaling |
Architectural Key Facts
- •Cloud-native relational engine delivering up to 5x throughput of standard MySQL and 3x of standard PostgreSQL.
- •Decouples compute from storage; data is replicated 6 ways across 3 Availability Zones with a 4-of-6 write quorum.
- •Up to 15 auto-scaling read replicas share the same underlying distributed 10 GB segment storage fleet with sub-10ms lag.
- •Aurora Serverless v2 scales compute instantly in fine-grained 0.5 ACU increments without dropping client connections.
- •Aurora Global Database replicates physical redo logs to secondary regions with <1 second replication latency.
Production Limits & Quotas
| Item / Dimension | Limit / Quota |
|---|---|
| Max database volume size | 128 TiB with automatic self-healing and growth |
| Max read replicas per cluster | 15 replicas |
| Aurora Serverless v2 max ACU | Up to 128 ACUs (256 GB RAM) |
Amazon DynamoDB
AWSKey Commands & CLI Operations
| Command | Description | |
|---|---|---|
| aws dynamodb create-table --table-name Users --attribute-definitions AttributeName=UserId,AttributeType=S --key-schema AttributeName=UserId,KeyType=HASH --billing-mode PAY_PER_REQUEST | Create On-Demand NoSQL table | |
| aws dynamodb put-item --table-name Users --item '{"UserId":{"S":"usr_100"},"Email":{"S":"dev@cloud.com"}}' | Insert document item | |
| aws dynamodb query --table-name Orders --key-condition-expression 'CustomerId = :c AND OrderDate >= :d' --expression-attribute-values '{":c":{"S":"cust_1"},":d":{"S":"2026-01-01"}}' | Query with partition and sort key | |
| aws dynamodb update-continuous-backups --table-name Users --point-in-time-recovery-specification PointInTimeRecoveryEnabled=true | Enable continuous PITR backups |
Architectural Key Facts
- •Fully managed serverless NoSQL key-value and document database delivering consistent single-digit millisecond latency.
- •Partition keys are hashed with MD5 to distribute data across internal partitions; automatic splits at 1,000 WCU / 3,000 RCU / 10 GB.
- •Global Secondary Indexes (GSIs) allow querying by alternative partition/sort keys with asynchronous projection.
- •DynamoDB Streams captures an ordered 24-hour changelog (CDC) for event-driven triggers via AWS Lambda.
- •Global Tables provides multi-region active-active masterless replication with Last-Writer-Wins conflict resolution.
Production Limits & Quotas
| Item / Dimension | Limit / Quota |
|---|---|
| Max single item size | 400 KB (including attribute names and values) |
| Max operations per transaction | 100 items or 4 MB per `TransactWriteItems` |
| GSIs per table (default) | 20 GSIs (expandable to 500 via soft limit) |
Amazon ElastiCache
AWSKey Commands & CLI Operations
| Command | Description | |
|---|---|---|
| aws elasticache create-replication-group --replication-group-id cache-cluster --replication-group-description 'Redis Cache' --engine redis --cache-node-type cache.r6g.large --num-node-groups 3 --replicas-per-node-group 2 --automatic-failover-enabled | Create 3-shard Redis cluster with HA | |
| aws elasticache describe-replication-groups --replication-group-id cache-cluster | View primary and replica cluster endpoints | |
| redis-cli -h CLUSTER_ENDPOINT -p 6379 -c SET session:user_99 '{"role":"admin"}' EX 3600 | Set key with 1-hour TTL via cluster mode client |
Architectural Key Facts
- •In-memory data store service providing sub-millisecond response times for Redis OSS / Valkey and Memcached engines.
- •Cluster Mode Enabled partitions data across 16,384 hash slots using CRC16 hashing for linear horizontal scaling.
- •Multi-AZ deployment with automatic failover detects primary failure and promotes a read replica in under 60 seconds.
- •Cache-Aside (Lazy Loading) and Write-Through caching patterns prevent database overload and cache stampedes.
- •ElastiCache Global Datastore provides cross-region active-passive replication with <1 second latency.
Production Limits & Quotas
| Item / Dimension | Limit / Quota |
|---|---|
| Max shards per cluster | Up to 500 shards in Cluster Mode |
| Max nodes per cluster | Up to 500 nodes |
| Max memory per node | Up to 419 GiB on `cache.r6g.16xlarge` |
Azure SQL Database
AzureKey Commands & CLI Operations
| Command | Description | |
|---|---|---|
| az sql server create -g myrg -n sqlsrv-prod -u sqladmin -p 'ComplexP@ss1' -l eastus | Provision logical Azure SQL Server | |
| az sql db create -g myrg -s sqlsrv-prod -n appdb --service-objective GP_Gen5_4 | Create General Purpose 4-vCore database | |
| az sql db create -g myrg -s sqlsrv-prod -n scale-db --tier Hyperscale --capacity 4 | Create Hyperscale database auto-scaling to 100TB | |
| az sql db failover -g myrg -s sqlsrv-prod -n appdb | Trigger failover to secondary replica |
Architectural Key Facts
- •Fully managed relational database engine based on the latest stable enterprise Microsoft SQL Server.
- •Purchasing models: vCore-based (recommended for transparency) and DTU-based (bundled compute/storage).
- •Hyperscale tier features a decoupled multi-tier architecture supporting databases up to 100 TB with rapid backups and restore in minutes.
- •Serverless compute tier automatically scales compute based on workload demands and pauses during periods of inactivity to save costs.
- •Active Geo-Replication creates up to four readable secondary databases in different Azure regions for disaster recovery.
Production Limits & Quotas
| Item / Dimension | Limit / Quota |
|---|---|
| Max database size (Hyperscale) | 100 TB with rapid storage allocation |
| Max vCores | Up to 128 vCores |
| Point-in-time restore (PITR) retention | 1 to 35 days (up to 10 years with Long-Term Retention) |
Azure Cosmos DB
AzureKey Commands & CLI Operations
| Command | Description | |
|---|---|---|
| az cosmosdb create -g myrg -n cosmos-prod --capabilities EnableServerless --default-consistency-level Session | Create serverless Cosmos DB account | |
| az cosmosdb sql database create -g myrg -a cosmos-prod -n AppDB | Create SQL (Core) API database | |
| az cosmosdb sql container create -g myrg -a cosmos-prod -d AppDB -n Users --partition-key-path '/userId' --throughput 1000 | Create container with partition key and provisioned RU/s |
Architectural Key Facts
- •Globally distributed, multi-model NoSQL database service guaranteeing single-digit millisecond latency at the 99th percentile.
- •Throughput is measured in Request Units (RU/s); supports Provisioned Throughput, Autoscale RU/s, and Serverless pay-per-request.
- •Offers 5 distinct, well-defined consistency levels: Strong, Bounded Staleness, Session (default), Consistent Prefix, and Eventual.
- •Multi-region write capability allows writing to any deployed replica worldwide with automatic conflict resolution.
- •Multi-model support: SQL (Core) API, MongoDB API, Apache Cassandra API, Gremlin (Graph) API, and Table API.
Production Limits & Quotas
| Item / Dimension | Limit / Quota |
|---|---|
| Max single document item size | 2 MB |
| Max physical partition storage | 50 GB per physical partition |
| Max RU/s per physical partition | 10,000 RU/s |
Azure Cache for Redis
AzureKey Commands & CLI Operations
| Command | Description | |
|---|---|---|
| az redis create -g myrg -n prod-cache --sku Premium --vm-size P1 --enable-non-ssl-port false --shard-count 2 | Create clustered 2-shard Redis cache | |
| az redis show -g myrg -n prod-cache --query 'hostName' | Get Redis SSL connection hostname |
Architectural Key Facts
- •Fully managed in-memory data store based on the open-source Redis engine for low-latency, high-throughput caching.
- •Tiers: Basic (single node), Standard (two-node primary/replica with 99.9% SLA), Premium (clustering, persistence, VNet injection), and Enterprise.
- •Enterprise tiers run Redis Enterprise software featuring RedisBloom, RediSearch, and Active-Active Geo-Distribution (CRDTs).
- •Data persistence via RDB backup and append-only file (AOF) protects cached data across reboots.
- •Cluster mode horizontally shards data across multiple nodes to exceed single VM memory limits.
Production Limits & Quotas
| Item / Dimension | Limit / Quota |
|---|---|
| Max cache size (Enterprise) | Up to 2 TB RAM |
| Max client connections | Up to 100,000 concurrent connections |
| Max shards per cluster | Up to 10 shards in Premium; 30 in Enterprise |
Cloud SQL
Google CloudKey Commands & CLI Operations
| Command | Description | |
|---|---|---|
| gcloud sql instances create my-instance --tier=db-custom-2-7680 --region=us-central1 | Create instance | |
| gcloud sql databases create my-db --instance=my-instance | Create database | |
| gcloud sql backups create --instance=my-instance | Trigger on-demand backup | |
| gcloud sql export sql my-instance gs://bucket/export.sql --database=my-db | Export data to GCS |
Architectural Key Facts
- •Fully managed relational database service supporting MySQL, PostgreSQL, and SQL Server.
- •Automates backups, replication, patches, and capacity management.
- •Supports High Availability (HA) through synchronous replication to a standby zone.
- •Cross-region read replicas available for disaster recovery and scaling read workloads.
- •Cloud SQL Auth Proxy provides secure, IAM-authenticated connections.
Production Limits & Quotas
| Item / Dimension | Limit / Quota |
|---|---|
| Max storage capacity | 64 TB |
| Max concurrent connections | Depends on instance memory |
| Max instances per project | 100 (soft quota) |
Firestore
Google CloudKey Commands & CLI Operations
| Command | Description | |
|---|---|---|
| gcloud firestore databases create --region=us-central | Initialize Firestore database | |
| gcloud firestore indexes composite create --collection-group=users --field-config... | Create composite index | |
| gcloud firestore export gs://my-bucket | Export Firestore data |
Architectural Key Facts
- •Flexible, scalable NoSQL document database for web, mobile, and server development.
- •Successor to Datastore; strongly consistent globally.
- •Real-time updates and offline support for mobile/web apps.
- •Organized in collections and documents; documents can contain subcollections.
- •ACID transactions across multiple documents.
Production Limits & Quotas
| Item / Dimension | Limit / Quota |
|---|---|
| Max document size | 1 MB |
| Max writes per second per document | 1 |
| Max documents returned in a query | Depends on timeout/memory, pagination recommended |
| Max composite indexes per db | 200 |
Spanner
Google CloudKey Commands & CLI Operations
| Command | Description | |
|---|---|---|
| gcloud spanner instances create my-instance --config=regional-us-central1 --nodes=1 | Create Spanner instance | |
| gcloud spanner databases create my-db --instance=my-instance | Create a database | |
| gcloud spanner databases execute-sql my-db --instance=my-instance --sql='SELECT 1' | Execute SQL |
Architectural Key Facts
- •Fully managed, globally distributed, horizontally scalable relational database.
- •Provides ACID transactions and strong consistency globally using TrueTime.
- •Combines relational structure (schemas, SQL queries) with non-relational scale.
- •99.999% SLA for multi-region instances.
- •Interleaved tables help colocate related child rows with parent rows for extreme performance.
Production Limits & Quotas
| Item / Dimension | Limit / Quota |
|---|---|
| Max nodes per instance | No hard limit |
| Max database size | Virtually unlimited |
| Max databases per instance | 100 |
| Storage per node limit | 4 TB |