[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-aws-aws-lambda-managed-instances":3,"mdc-78yft7-key":35,"related-repo-aws-aws-lambda-managed-instances":1348,"related-org-aws-aws-lambda-managed-instances":1450},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":24,"repoUrl":25,"updatedAt":26,"license":27,"forks":28,"topics":29,"repo":30,"sourceUrl":33,"mdContent":34},"aws-lambda-managed-instances","migrate workloads to AWS Lambda Managed Instances","Evaluates, configures, and migrates workloads to AWS Lambda Managed Instances (LMI). Runs Lambda functions on EC2 instances in the user's account while AWS manages provisioning, patching, scaling, routing, and load balancing. Triggers when queries mention Lambda Managed Instances, LMI, capacity providers, multi-concurrent execution environments, EC2-backed Lambda, persistent Lambda instances, PerExecutionEnvironmentMaxConcurrency, CapacityProviderConfig, cold start elimination via dedicated instances, migrating standard Lambda to managed instances, or cost comparison between standard Lambda and LMI with Savings Plans or Reserved Instances.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},"aws","AWS (Amazon)","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Faws.png",[12,16,19,22],{"name":13,"slug":14,"type":15},"Modernization","modernization","tag",{"name":17,"slug":18,"type":15},"Deployment","deployment",{"name":20,"slug":21,"type":15},"Infrastructure","infrastructure",{"name":23,"slug":8,"type":15},"AWS",1822,"https:\u002F\u002Fgithub.com\u002Faws\u002Fagent-toolkit-for-aws","2026-07-12T08:44:33.665075",null,157,[],{"repoUrl":25,"stars":24,"forks":28,"topics":31,"description":32},[],"Official, AWS-supported MCP servers, skills, and plugins to help AI agents build on AWS","https:\u002F\u002Fgithub.com\u002Faws\u002Fagent-toolkit-for-aws\u002Ftree\u002FHEAD\u002Fskills\u002Fspecialized-skills\u002Fserverless-skills\u002Faws-lambda-managed-instances","---\nname: aws-lambda-managed-instances\ndescription: \"Evaluates, configures, and migrates workloads to AWS Lambda Managed Instances (LMI). Runs Lambda functions on EC2 instances in the user's account while AWS manages provisioning, patching, scaling, routing, and load balancing. Triggers when queries mention Lambda Managed Instances, LMI, capacity providers, multi-concurrent execution environments, EC2-backed Lambda, persistent Lambda instances, PerExecutionEnvironmentMaxConcurrency, CapacityProviderConfig, cold start elimination via dedicated instances, migrating standard Lambda to managed instances, or cost comparison between standard Lambda and LMI with Savings Plans or Reserved Instances.\"\nversion: 1\n---\n\n# AWS Lambda Managed Instances (LMI)\n\nRuns Lambda functions on EC2 instances in the user's account while AWS manages provisioning, patching, scaling, routing, and load balancing. Combines Lambda's developer experience with EC2's pricing and hardware options.\n\n**Works best with** the [AWS MCP server](https:\u002F\u002Fdocs.aws.amazon.com\u002Faws-mcp\u002F) for sandboxed CLI execution and audit logging. All guidance also works with standard AWS CLI or SAM CLI.\n\n**Note:** Confirm regional availability, quotas, and instance type offerings against current AWS documentation before production deployment.\n\n## Quick Decision: Is LMI Right for This Workload?\n\n| Signal | LMI is a strong fit | Standard Lambda is better |\n|--------|---------------------|---------------------------|\n| Traffic | Steady, predictable, 50M+ req\u002Fmo | Bursty, unpredictable, long periods of no traffic |\n| Cost | Duration-heavy spend at scale | Low or sporadic invocations |\n| Cold starts | Unacceptable (LMI eliminates for provisioned capacity) | Tolerable |\n| Compute | Latest CPUs, specific families, high network bandwidth, GPU requirements | Standard Lambda memory\u002FCPU sufficient |\n| Isolation | Dedicated EC2 instances in your account, full VPC control | Shared Firecracker micro-VMs acceptable |\n| Scale-to-zero | Does not scale to zero but can create custom schedules with AWS provided solutions | Required (pay nothing when idle) |\n| Code readiness | Thread-safe (Node.js\u002FJava\u002F.NET) or any Python code | Non-thread-safe code, expensive to change |\n\n## Routing\n\nRead ONLY the single reference file that matches the user's task. Do not preload multiple references.\n\n| User need | Action |\n|-----------|--------|\n| Cost comparison, pricing analysis, Savings Plans, Reserved Instances | Read [cost-comparison.md](references\u002Fcost-comparison.md) |\n| Instance types, memory sizing, vCPU ratios, scaling tuning, capacity provider config | Read [configuration-guide.md](references\u002Fconfiguration-guide.md) |\n| Thread safety, concurrency model, code review checklist, multi-concurrency readiness | Read [thread-safety.md](references\u002Fthread-safety.md) |\n| Before\u002Fafter code examples, runtime-specific migration, connection pooling | Read [migration-patterns.md](references\u002Fmigration-patterns.md) |\n| IAM roles, VPC setup, CLI commands, SAM template, CDK example | Read [infrastructure-setup.md](references\u002Finfrastructure-setup.md) |\n| Errors, throttling, debugging, stuck deployments | Read [troubleshooting.md](references\u002Ftroubleshooting.md) |\n\n**Troubleshooting quick facts** (always mention when diagnosing issues):\n\n- Capacity provider stuck in CREATING → most common cause is **private subnets missing a NAT gateway route** (instances need outbound internet for image pull and Lambda service communication)\n- Function not scaling → check that a **version is published** (PublishToLatestPublished: true)\n- Memory errors → LMI minimum is **2048 MB**\n\n## Workflow\n\n### Step 1: Assess the Workload\n\nGather these signals before recommending:\n\n1. **Traffic pattern**: Steady vs bursty? Requests per second?\n2. **Current costs**: Monthly Lambda spend? Existing Savings Plans?\n3. **Runtime**: Node.js, Java, .NET, or Python?\n4. **Memory\u002FCPU**: How much memory? CPU-bound or I\u002FO-bound?\n5. **Execution duration**: Average and P99?\n6. **Concurrency readiness**: Thread safety? Shared `\u002Ftmp` paths? Per-invocation DB connections?\n7. **VPC**: Already in a VPC? Private resource access needed?\n\nWhen recommending LMI, ALWAYS mention: minimum 3 execution environments for AZ resiliency (cannot go below 3 in production).\n\n### Step 2: Build the Cost Comparison\n\nREQUIRED: Present a cost comparison before recommending LMI.\n\nRule of thumb: LMI becomes cost-competitive at 50-100M+ req\u002Fmonth with steady traffic. Use the [LMI Pricing Calculator](https:\u002F\u002Faws-samples.github.io\u002Fsample-aws-lambda-managed-instances\u002F) for accurate comparisons.\n\n### Step 3: Configure the Deployment\n\n- **Instance families** (400+ types, .large and up): C-series (compute), M-series (general), R-series (memory). ARM (Graviton) for best price-performance.\n- **When using Graviton instances, MUST set `Architectures: [arm64]`** in the function configuration to match.\n- **Memory-to-vCPU ratios**: 2:1 (compute), 4:1 (general, default), 8:1 (memory). Min 2 GB, max 32 GB.\n- **Multi-concurrency per-vCPU maximums**: Node.js 64, Java 32, .NET 32, Python 16. These are system caps — the actual setting is PerExecutionEnvironmentMaxConcurrency (per execution environment, not per vCPU).\n- **For I\u002FO-bound workloads**: use the runtime default or higher PerExecutionEnvironmentMaxConcurrency (e.g., 10 for Node.js) since each request uses minimal CPU while waiting on network.\n- **For CPU-bound workloads**: set PerExecutionEnvironmentMaxConcurrency to 1-2 per vCPU since each request saturates CPU.\n- **Scaling**: MinExecutionEnvironments (default 3), MaxVCpuCount (optional, default 400 — set explicitly as best practice), TargetResourceUtilization.\n\n### Step 4: Migrate the Code\n\nReview code for concurrency safety. LMI runs multiple invocations concurrently per execution environment:\n\n- **Python**: Process-based isolation — globals are NOT shared. No thread-safety changes needed. Focus on `\u002Ftmp` conflicts and memory sizing.\n- **Node.js**: Worker threads — globals shared within a worker. Requires async safety.\n- **Java\u002F.NET**: OS threads\u002FTasks — handler shared across threads. Requires full thread safety.\n\n### Step 5: Set Up Infrastructure\n\n1. Create two IAM roles: execution role (for the function) and operator role (for capacity provider EC2 management)\n2. Configure VPC with subnets across 3+ AZs\n3. Create capacity provider with VPC config and scaling limits\n4. Create or update function with capacity provider attachment\n5. Publish a version (triggers instance provisioning)\n\n### Step 6: Validate and Cut Over\n\n1. Deploy to a non-production environment first\n2. Monitor CloudWatch: CPU utilization, memory, concurrency, throttle rate\n3. Gradual traffic shift with weighted aliases (10% → 50% → 100%)\n4. Compare costs after 1-2 weeks of production data\n5. Decommission standard Lambda once stable\n\n## Best Practices\n\n### Pricing (always mention when discussing costs)\n\n- **Three components**: EC2 instance hours + 15% management fee + $0.20\u002F1M requests\n- **Savings Plans**: Compute Savings Plans apply to the EC2 portion (up to 60-72% discount)\n- **The 15% fee** is charged on top of EC2 cost for AWS managing provisioning, patching, scaling, lifecycle\n\n### Scaling (always mention when discussing scaling or traffic)\n\n- LMI absorbs a 50% traffic spike immediately and **doubles capacity within 5 minutes** — if traffic more than doubles faster, requests throttle\n- Standard Lambda bursts to 3000 instantly — LMI cannot match this\n- **Pre-warm** with MinExecutionEnvironments before known spikes\n- **MaxVCpuCount** (default 400) — set explicitly as a cost ceiling\n- **Shape**: Reduce MinExecutionEnvironments to lower capacity during off-hours (minimum 3 for AZ resiliency)\n\n### Instance Sizing\n\n- **1 vCPU + 1 GB reserved per instance** for OS overhead (not available to your function)\n- Usable capacity = total - overhead\n\n### Configuration\n\n- Start with 4:1 ratio and runtime default concurrency\n- Use ARM (Graviton) unless x86 dependencies exist\n- Let Lambda choose instance types unless specific hardware needed\n- Set MaxVCpuCount to control cost ceiling\n- Never set MinExecutionEnvironments below 3 (breaks AZ resiliency)\n\n### Migration\n\n- Start with I\u002FO-heavy functions (benefit most from multi-concurrency)\n- Review code for concurrency safety before attaching to capacity provider\n- Use weighted aliases for gradual traffic shift\n- Include request IDs in all log statements\n- Initialize DB pools and SDK clients outside the handler\n\n### Operations\n\n- Set CloudWatch alarms on throttle rate > 1% and CPU > 80%\n- Plan for 14-day instance rotation (automatic)\n- Never manually terminate LMI EC2 instances (delete the capacity provider instead)\n- Always publish a version — unpublished functions cannot run on LMI\n\n## Limits Quick Reference\n\n| Resource | Limit |\n|----------|-------|\n| Memory | 2 GB min, 32 GB max |\n| Execution environments | 3 minimum (MinExecutionEnvironments, AZ resiliency) |\n| Instance lifespan | 14 days (auto-replaced) |\n| Concurrency\u002FvCPU | 64 (Node.js), 32 (Java\u002F.NET), 16 (Python) |\n| Runtimes | Node.js 22+, Java 21+, .NET 8+, Python 3.13+, Rust (provided.al2023) |\n| Instance families | C, M, R (.large and up) |\n| Scaling | Burst headroom equals unused capacity from TargetResourceUtilization; new instances launch within minutes |\n\n## Security Considerations\n\n- **Operator role scoping**: Add `aws:SourceAccount` and `aws:SourceArn` conditions to trust policies to prevent confused deputy attacks.\n- **VPC egress**: Scope security group egress to VPC endpoint security groups or AWS prefix lists rather than 0.0.0.0\u002F0.\n- **Credentials**: Use AWS Secrets Manager or Parameter Store for database credentials — never environment variables for secrets.\n- **Encryption**: Enable SQS SSE, CloudWatch Logs encryption (KMS), and S3 default encryption for any data at rest.\n- **Logging**: Set CloudWatch Log group retention policies. Avoid logging PII or credentials. Enable CloudTrail data events for Lambda.\n- **Instance rotation**: The 14-day automatic rotation ensures security patches are applied without manual intervention.\n- **References**: [Lambda Security Best Practices](https:\u002F\u002Fdocs.aws.amazon.com\u002Flambda\u002Flatest\u002Fdg\u002Flambda-security.html), [IAM Best Practices](https:\u002F\u002Fdocs.aws.amazon.com\u002FIAM\u002Flatest\u002FUserGuide\u002Fbest-practices.html)\n\n## Files\n\n| File | Content |\n|------|---------|\n| [cost-comparison.md](references\u002Fcost-comparison.md) | Pricing analysis, break-even calculations, Savings Plans\u002FRI impact |\n| [configuration-guide.md](references\u002Fconfiguration-guide.md) | Instance selection, memory ratios, scaling tuning, capacity provider config |\n| [thread-safety.md](references\u002Fthread-safety.md) | Concurrency model per runtime, code review checklist, Powertools compatibility |\n| [migration-patterns.md](references\u002Fmigration-patterns.md) | Before\u002Fafter code by runtime, connection pooling, gradual cutover |\n| [infrastructure-setup.md](references\u002Finfrastructure-setup.md) | IAM roles, VPC setup, SAM templates, CLI commands |\n| [troubleshooting.md](references\u002Ftroubleshooting.md) | Common errors, throttling, debugging, stuck deployments |\n",{"data":36,"body":38},{"name":4,"description":6,"version":37},1,{"type":39,"children":40},"root",[41,50,56,78,88,95,254,260,265,396,406,445,451,458,463,546,551,557,562,576,582,661,667,672,712,718,746,752,780,786,792,825,831,881,887,905,911,939,945,973,979,1002,1008,1119,1125,1230,1236],{"type":42,"tag":43,"props":44,"children":46},"element","h1",{"id":45},"aws-lambda-managed-instances-lmi",[47],{"type":48,"value":49},"text","AWS Lambda Managed Instances (LMI)",{"type":42,"tag":51,"props":52,"children":53},"p",{},[54],{"type":48,"value":55},"Runs Lambda functions on EC2 instances in the user's account while AWS manages provisioning, patching, scaling, routing, and load balancing. Combines Lambda's developer experience with EC2's pricing and hardware options.",{"type":42,"tag":51,"props":57,"children":58},{},[59,65,67,76],{"type":42,"tag":60,"props":61,"children":62},"strong",{},[63],{"type":48,"value":64},"Works best with",{"type":48,"value":66}," the ",{"type":42,"tag":68,"props":69,"children":73},"a",{"href":70,"rel":71},"https:\u002F\u002Fdocs.aws.amazon.com\u002Faws-mcp\u002F",[72],"nofollow",[74],{"type":48,"value":75},"AWS MCP server",{"type":48,"value":77}," for sandboxed CLI execution and audit logging. All guidance also works with standard AWS CLI or SAM CLI.",{"type":42,"tag":51,"props":79,"children":80},{},[81,86],{"type":42,"tag":60,"props":82,"children":83},{},[84],{"type":48,"value":85},"Note:",{"type":48,"value":87}," Confirm regional availability, quotas, and instance type offerings against current AWS documentation before production deployment.",{"type":42,"tag":89,"props":90,"children":92},"h2",{"id":91},"quick-decision-is-lmi-right-for-this-workload",[93],{"type":48,"value":94},"Quick Decision: Is LMI Right for This Workload?",{"type":42,"tag":96,"props":97,"children":98},"table",{},[99,123],{"type":42,"tag":100,"props":101,"children":102},"thead",{},[103],{"type":42,"tag":104,"props":105,"children":106},"tr",{},[107,113,118],{"type":42,"tag":108,"props":109,"children":110},"th",{},[111],{"type":48,"value":112},"Signal",{"type":42,"tag":108,"props":114,"children":115},{},[116],{"type":48,"value":117},"LMI is a strong fit",{"type":42,"tag":108,"props":119,"children":120},{},[121],{"type":48,"value":122},"Standard Lambda is better",{"type":42,"tag":124,"props":125,"children":126},"tbody",{},[127,146,164,182,200,218,236],{"type":42,"tag":104,"props":128,"children":129},{},[130,136,141],{"type":42,"tag":131,"props":132,"children":133},"td",{},[134],{"type":48,"value":135},"Traffic",{"type":42,"tag":131,"props":137,"children":138},{},[139],{"type":48,"value":140},"Steady, predictable, 50M+ req\u002Fmo",{"type":42,"tag":131,"props":142,"children":143},{},[144],{"type":48,"value":145},"Bursty, unpredictable, long periods of no traffic",{"type":42,"tag":104,"props":147,"children":148},{},[149,154,159],{"type":42,"tag":131,"props":150,"children":151},{},[152],{"type":48,"value":153},"Cost",{"type":42,"tag":131,"props":155,"children":156},{},[157],{"type":48,"value":158},"Duration-heavy spend at scale",{"type":42,"tag":131,"props":160,"children":161},{},[162],{"type":48,"value":163},"Low or sporadic invocations",{"type":42,"tag":104,"props":165,"children":166},{},[167,172,177],{"type":42,"tag":131,"props":168,"children":169},{},[170],{"type":48,"value":171},"Cold starts",{"type":42,"tag":131,"props":173,"children":174},{},[175],{"type":48,"value":176},"Unacceptable (LMI eliminates for provisioned capacity)",{"type":42,"tag":131,"props":178,"children":179},{},[180],{"type":48,"value":181},"Tolerable",{"type":42,"tag":104,"props":183,"children":184},{},[185,190,195],{"type":42,"tag":131,"props":186,"children":187},{},[188],{"type":48,"value":189},"Compute",{"type":42,"tag":131,"props":191,"children":192},{},[193],{"type":48,"value":194},"Latest CPUs, specific families, high network bandwidth, GPU requirements",{"type":42,"tag":131,"props":196,"children":197},{},[198],{"type":48,"value":199},"Standard Lambda memory\u002FCPU sufficient",{"type":42,"tag":104,"props":201,"children":202},{},[203,208,213],{"type":42,"tag":131,"props":204,"children":205},{},[206],{"type":48,"value":207},"Isolation",{"type":42,"tag":131,"props":209,"children":210},{},[211],{"type":48,"value":212},"Dedicated EC2 instances in your account, full VPC control",{"type":42,"tag":131,"props":214,"children":215},{},[216],{"type":48,"value":217},"Shared Firecracker micro-VMs acceptable",{"type":42,"tag":104,"props":219,"children":220},{},[221,226,231],{"type":42,"tag":131,"props":222,"children":223},{},[224],{"type":48,"value":225},"Scale-to-zero",{"type":42,"tag":131,"props":227,"children":228},{},[229],{"type":48,"value":230},"Does not scale to zero but can create custom schedules with AWS provided solutions",{"type":42,"tag":131,"props":232,"children":233},{},[234],{"type":48,"value":235},"Required (pay nothing when idle)",{"type":42,"tag":104,"props":237,"children":238},{},[239,244,249],{"type":42,"tag":131,"props":240,"children":241},{},[242],{"type":48,"value":243},"Code readiness",{"type":42,"tag":131,"props":245,"children":246},{},[247],{"type":48,"value":248},"Thread-safe (Node.js\u002FJava\u002F.NET) or any Python code",{"type":42,"tag":131,"props":250,"children":251},{},[252],{"type":48,"value":253},"Non-thread-safe code, expensive to change",{"type":42,"tag":89,"props":255,"children":257},{"id":256},"routing",[258],{"type":48,"value":259},"Routing",{"type":42,"tag":51,"props":261,"children":262},{},[263],{"type":48,"value":264},"Read ONLY the single reference file that matches the user's task. Do not preload multiple references.",{"type":42,"tag":96,"props":266,"children":267},{},[268,284],{"type":42,"tag":100,"props":269,"children":270},{},[271],{"type":42,"tag":104,"props":272,"children":273},{},[274,279],{"type":42,"tag":108,"props":275,"children":276},{},[277],{"type":48,"value":278},"User need",{"type":42,"tag":108,"props":280,"children":281},{},[282],{"type":48,"value":283},"Action",{"type":42,"tag":124,"props":285,"children":286},{},[287,306,324,342,360,378],{"type":42,"tag":104,"props":288,"children":289},{},[290,295],{"type":42,"tag":131,"props":291,"children":292},{},[293],{"type":48,"value":294},"Cost comparison, pricing analysis, Savings Plans, Reserved Instances",{"type":42,"tag":131,"props":296,"children":297},{},[298,300],{"type":48,"value":299},"Read ",{"type":42,"tag":68,"props":301,"children":303},{"href":302},"references\u002Fcost-comparison.md",[304],{"type":48,"value":305},"cost-comparison.md",{"type":42,"tag":104,"props":307,"children":308},{},[309,314],{"type":42,"tag":131,"props":310,"children":311},{},[312],{"type":48,"value":313},"Instance types, memory sizing, vCPU ratios, scaling tuning, capacity provider config",{"type":42,"tag":131,"props":315,"children":316},{},[317,318],{"type":48,"value":299},{"type":42,"tag":68,"props":319,"children":321},{"href":320},"references\u002Fconfiguration-guide.md",[322],{"type":48,"value":323},"configuration-guide.md",{"type":42,"tag":104,"props":325,"children":326},{},[327,332],{"type":42,"tag":131,"props":328,"children":329},{},[330],{"type":48,"value":331},"Thread safety, concurrency model, code review checklist, multi-concurrency readiness",{"type":42,"tag":131,"props":333,"children":334},{},[335,336],{"type":48,"value":299},{"type":42,"tag":68,"props":337,"children":339},{"href":338},"references\u002Fthread-safety.md",[340],{"type":48,"value":341},"thread-safety.md",{"type":42,"tag":104,"props":343,"children":344},{},[345,350],{"type":42,"tag":131,"props":346,"children":347},{},[348],{"type":48,"value":349},"Before\u002Fafter code examples, runtime-specific migration, connection pooling",{"type":42,"tag":131,"props":351,"children":352},{},[353,354],{"type":48,"value":299},{"type":42,"tag":68,"props":355,"children":357},{"href":356},"references\u002Fmigration-patterns.md",[358],{"type":48,"value":359},"migration-patterns.md",{"type":42,"tag":104,"props":361,"children":362},{},[363,368],{"type":42,"tag":131,"props":364,"children":365},{},[366],{"type":48,"value":367},"IAM roles, VPC setup, CLI commands, SAM template, CDK example",{"type":42,"tag":131,"props":369,"children":370},{},[371,372],{"type":48,"value":299},{"type":42,"tag":68,"props":373,"children":375},{"href":374},"references\u002Finfrastructure-setup.md",[376],{"type":48,"value":377},"infrastructure-setup.md",{"type":42,"tag":104,"props":379,"children":380},{},[381,386],{"type":42,"tag":131,"props":382,"children":383},{},[384],{"type":48,"value":385},"Errors, throttling, debugging, stuck deployments",{"type":42,"tag":131,"props":387,"children":388},{},[389,390],{"type":48,"value":299},{"type":42,"tag":68,"props":391,"children":393},{"href":392},"references\u002Ftroubleshooting.md",[394],{"type":48,"value":395},"troubleshooting.md",{"type":42,"tag":51,"props":397,"children":398},{},[399,404],{"type":42,"tag":60,"props":400,"children":401},{},[402],{"type":48,"value":403},"Troubleshooting quick facts",{"type":48,"value":405}," (always mention when diagnosing issues):",{"type":42,"tag":407,"props":408,"children":409},"ul",{},[410,423,435],{"type":42,"tag":411,"props":412,"children":413},"li",{},[414,416,421],{"type":48,"value":415},"Capacity provider stuck in CREATING → most common cause is ",{"type":42,"tag":60,"props":417,"children":418},{},[419],{"type":48,"value":420},"private subnets missing a NAT gateway route",{"type":48,"value":422}," (instances need outbound internet for image pull and Lambda service communication)",{"type":42,"tag":411,"props":424,"children":425},{},[426,428,433],{"type":48,"value":427},"Function not scaling → check that a ",{"type":42,"tag":60,"props":429,"children":430},{},[431],{"type":48,"value":432},"version is published",{"type":48,"value":434}," (PublishToLatestPublished: true)",{"type":42,"tag":411,"props":436,"children":437},{},[438,440],{"type":48,"value":439},"Memory errors → LMI minimum is ",{"type":42,"tag":60,"props":441,"children":442},{},[443],{"type":48,"value":444},"2048 MB",{"type":42,"tag":89,"props":446,"children":448},{"id":447},"workflow",[449],{"type":48,"value":450},"Workflow",{"type":42,"tag":452,"props":453,"children":455},"h3",{"id":454},"step-1-assess-the-workload",[456],{"type":48,"value":457},"Step 1: Assess the Workload",{"type":42,"tag":51,"props":459,"children":460},{},[461],{"type":48,"value":462},"Gather these signals before recommending:",{"type":42,"tag":464,"props":465,"children":466},"ol",{},[467,477,487,497,507,517,536],{"type":42,"tag":411,"props":468,"children":469},{},[470,475],{"type":42,"tag":60,"props":471,"children":472},{},[473],{"type":48,"value":474},"Traffic pattern",{"type":48,"value":476},": Steady vs bursty? Requests per second?",{"type":42,"tag":411,"props":478,"children":479},{},[480,485],{"type":42,"tag":60,"props":481,"children":482},{},[483],{"type":48,"value":484},"Current costs",{"type":48,"value":486},": Monthly Lambda spend? Existing Savings Plans?",{"type":42,"tag":411,"props":488,"children":489},{},[490,495],{"type":42,"tag":60,"props":491,"children":492},{},[493],{"type":48,"value":494},"Runtime",{"type":48,"value":496},": Node.js, Java, .NET, or Python?",{"type":42,"tag":411,"props":498,"children":499},{},[500,505],{"type":42,"tag":60,"props":501,"children":502},{},[503],{"type":48,"value":504},"Memory\u002FCPU",{"type":48,"value":506},": How much memory? CPU-bound or I\u002FO-bound?",{"type":42,"tag":411,"props":508,"children":509},{},[510,515],{"type":42,"tag":60,"props":511,"children":512},{},[513],{"type":48,"value":514},"Execution duration",{"type":48,"value":516},": Average and P99?",{"type":42,"tag":411,"props":518,"children":519},{},[520,525,527,534],{"type":42,"tag":60,"props":521,"children":522},{},[523],{"type":48,"value":524},"Concurrency readiness",{"type":48,"value":526},": Thread safety? Shared ",{"type":42,"tag":528,"props":529,"children":531},"code",{"className":530},[],[532],{"type":48,"value":533},"\u002Ftmp",{"type":48,"value":535}," paths? Per-invocation DB connections?",{"type":42,"tag":411,"props":537,"children":538},{},[539,544],{"type":42,"tag":60,"props":540,"children":541},{},[542],{"type":48,"value":543},"VPC",{"type":48,"value":545},": Already in a VPC? Private resource access needed?",{"type":42,"tag":51,"props":547,"children":548},{},[549],{"type":48,"value":550},"When recommending LMI, ALWAYS mention: minimum 3 execution environments for AZ resiliency (cannot go below 3 in production).",{"type":42,"tag":452,"props":552,"children":554},{"id":553},"step-2-build-the-cost-comparison",[555],{"type":48,"value":556},"Step 2: Build the Cost Comparison",{"type":42,"tag":51,"props":558,"children":559},{},[560],{"type":48,"value":561},"REQUIRED: Present a cost comparison before recommending LMI.",{"type":42,"tag":51,"props":563,"children":564},{},[565,567,574],{"type":48,"value":566},"Rule of thumb: LMI becomes cost-competitive at 50-100M+ req\u002Fmonth with steady traffic. Use the ",{"type":42,"tag":68,"props":568,"children":571},{"href":569,"rel":570},"https:\u002F\u002Faws-samples.github.io\u002Fsample-aws-lambda-managed-instances\u002F",[72],[572],{"type":48,"value":573},"LMI Pricing Calculator",{"type":48,"value":575}," for accurate comparisons.",{"type":42,"tag":452,"props":577,"children":579},{"id":578},"step-3-configure-the-deployment",[580],{"type":48,"value":581},"Step 3: Configure the Deployment",{"type":42,"tag":407,"props":583,"children":584},{},[585,595,611,621,631,641,651],{"type":42,"tag":411,"props":586,"children":587},{},[588,593],{"type":42,"tag":60,"props":589,"children":590},{},[591],{"type":48,"value":592},"Instance families",{"type":48,"value":594}," (400+ types, .large and up): C-series (compute), M-series (general), R-series (memory). ARM (Graviton) for best price-performance.",{"type":42,"tag":411,"props":596,"children":597},{},[598,609],{"type":42,"tag":60,"props":599,"children":600},{},[601,603],{"type":48,"value":602},"When using Graviton instances, MUST set ",{"type":42,"tag":528,"props":604,"children":606},{"className":605},[],[607],{"type":48,"value":608},"Architectures: [arm64]",{"type":48,"value":610}," in the function configuration to match.",{"type":42,"tag":411,"props":612,"children":613},{},[614,619],{"type":42,"tag":60,"props":615,"children":616},{},[617],{"type":48,"value":618},"Memory-to-vCPU ratios",{"type":48,"value":620},": 2:1 (compute), 4:1 (general, default), 8:1 (memory). Min 2 GB, max 32 GB.",{"type":42,"tag":411,"props":622,"children":623},{},[624,629],{"type":42,"tag":60,"props":625,"children":626},{},[627],{"type":48,"value":628},"Multi-concurrency per-vCPU maximums",{"type":48,"value":630},": Node.js 64, Java 32, .NET 32, Python 16. These are system caps — the actual setting is PerExecutionEnvironmentMaxConcurrency (per execution environment, not per vCPU).",{"type":42,"tag":411,"props":632,"children":633},{},[634,639],{"type":42,"tag":60,"props":635,"children":636},{},[637],{"type":48,"value":638},"For I\u002FO-bound workloads",{"type":48,"value":640},": use the runtime default or higher PerExecutionEnvironmentMaxConcurrency (e.g., 10 for Node.js) since each request uses minimal CPU while waiting on network.",{"type":42,"tag":411,"props":642,"children":643},{},[644,649],{"type":42,"tag":60,"props":645,"children":646},{},[647],{"type":48,"value":648},"For CPU-bound workloads",{"type":48,"value":650},": set PerExecutionEnvironmentMaxConcurrency to 1-2 per vCPU since each request saturates CPU.",{"type":42,"tag":411,"props":652,"children":653},{},[654,659],{"type":42,"tag":60,"props":655,"children":656},{},[657],{"type":48,"value":658},"Scaling",{"type":48,"value":660},": MinExecutionEnvironments (default 3), MaxVCpuCount (optional, default 400 — set explicitly as best practice), TargetResourceUtilization.",{"type":42,"tag":452,"props":662,"children":664},{"id":663},"step-4-migrate-the-code",[665],{"type":48,"value":666},"Step 4: Migrate the Code",{"type":42,"tag":51,"props":668,"children":669},{},[670],{"type":48,"value":671},"Review code for concurrency safety. LMI runs multiple invocations concurrently per execution environment:",{"type":42,"tag":407,"props":673,"children":674},{},[675,692,702],{"type":42,"tag":411,"props":676,"children":677},{},[678,683,685,690],{"type":42,"tag":60,"props":679,"children":680},{},[681],{"type":48,"value":682},"Python",{"type":48,"value":684},": Process-based isolation — globals are NOT shared. No thread-safety changes needed. Focus on ",{"type":42,"tag":528,"props":686,"children":688},{"className":687},[],[689],{"type":48,"value":533},{"type":48,"value":691}," conflicts and memory sizing.",{"type":42,"tag":411,"props":693,"children":694},{},[695,700],{"type":42,"tag":60,"props":696,"children":697},{},[698],{"type":48,"value":699},"Node.js",{"type":48,"value":701},": Worker threads — globals shared within a worker. Requires async safety.",{"type":42,"tag":411,"props":703,"children":704},{},[705,710],{"type":42,"tag":60,"props":706,"children":707},{},[708],{"type":48,"value":709},"Java\u002F.NET",{"type":48,"value":711},": OS threads\u002FTasks — handler shared across threads. Requires full thread safety.",{"type":42,"tag":452,"props":713,"children":715},{"id":714},"step-5-set-up-infrastructure",[716],{"type":48,"value":717},"Step 5: Set Up Infrastructure",{"type":42,"tag":464,"props":719,"children":720},{},[721,726,731,736,741],{"type":42,"tag":411,"props":722,"children":723},{},[724],{"type":48,"value":725},"Create two IAM roles: execution role (for the function) and operator role (for capacity provider EC2 management)",{"type":42,"tag":411,"props":727,"children":728},{},[729],{"type":48,"value":730},"Configure VPC with subnets across 3+ AZs",{"type":42,"tag":411,"props":732,"children":733},{},[734],{"type":48,"value":735},"Create capacity provider with VPC config and scaling limits",{"type":42,"tag":411,"props":737,"children":738},{},[739],{"type":48,"value":740},"Create or update function with capacity provider attachment",{"type":42,"tag":411,"props":742,"children":743},{},[744],{"type":48,"value":745},"Publish a version (triggers instance provisioning)",{"type":42,"tag":452,"props":747,"children":749},{"id":748},"step-6-validate-and-cut-over",[750],{"type":48,"value":751},"Step 6: Validate and Cut Over",{"type":42,"tag":464,"props":753,"children":754},{},[755,760,765,770,775],{"type":42,"tag":411,"props":756,"children":757},{},[758],{"type":48,"value":759},"Deploy to a non-production environment first",{"type":42,"tag":411,"props":761,"children":762},{},[763],{"type":48,"value":764},"Monitor CloudWatch: CPU utilization, memory, concurrency, throttle rate",{"type":42,"tag":411,"props":766,"children":767},{},[768],{"type":48,"value":769},"Gradual traffic shift with weighted aliases (10% → 50% → 100%)",{"type":42,"tag":411,"props":771,"children":772},{},[773],{"type":48,"value":774},"Compare costs after 1-2 weeks of production data",{"type":42,"tag":411,"props":776,"children":777},{},[778],{"type":48,"value":779},"Decommission standard Lambda once stable",{"type":42,"tag":89,"props":781,"children":783},{"id":782},"best-practices",[784],{"type":48,"value":785},"Best Practices",{"type":42,"tag":452,"props":787,"children":789},{"id":788},"pricing-always-mention-when-discussing-costs",[790],{"type":48,"value":791},"Pricing (always mention when discussing costs)",{"type":42,"tag":407,"props":793,"children":794},{},[795,805,815],{"type":42,"tag":411,"props":796,"children":797},{},[798,803],{"type":42,"tag":60,"props":799,"children":800},{},[801],{"type":48,"value":802},"Three components",{"type":48,"value":804},": EC2 instance hours + 15% management fee + $0.20\u002F1M requests",{"type":42,"tag":411,"props":806,"children":807},{},[808,813],{"type":42,"tag":60,"props":809,"children":810},{},[811],{"type":48,"value":812},"Savings Plans",{"type":48,"value":814},": Compute Savings Plans apply to the EC2 portion (up to 60-72% discount)",{"type":42,"tag":411,"props":816,"children":817},{},[818,823],{"type":42,"tag":60,"props":819,"children":820},{},[821],{"type":48,"value":822},"The 15% fee",{"type":48,"value":824}," is charged on top of EC2 cost for AWS managing provisioning, patching, scaling, lifecycle",{"type":42,"tag":452,"props":826,"children":828},{"id":827},"scaling-always-mention-when-discussing-scaling-or-traffic",[829],{"type":48,"value":830},"Scaling (always mention when discussing scaling or traffic)",{"type":42,"tag":407,"props":832,"children":833},{},[834,846,851,861,871],{"type":42,"tag":411,"props":835,"children":836},{},[837,839,844],{"type":48,"value":838},"LMI absorbs a 50% traffic spike immediately and ",{"type":42,"tag":60,"props":840,"children":841},{},[842],{"type":48,"value":843},"doubles capacity within 5 minutes",{"type":48,"value":845}," — if traffic more than doubles faster, requests throttle",{"type":42,"tag":411,"props":847,"children":848},{},[849],{"type":48,"value":850},"Standard Lambda bursts to 3000 instantly — LMI cannot match this",{"type":42,"tag":411,"props":852,"children":853},{},[854,859],{"type":42,"tag":60,"props":855,"children":856},{},[857],{"type":48,"value":858},"Pre-warm",{"type":48,"value":860}," with MinExecutionEnvironments before known spikes",{"type":42,"tag":411,"props":862,"children":863},{},[864,869],{"type":42,"tag":60,"props":865,"children":866},{},[867],{"type":48,"value":868},"MaxVCpuCount",{"type":48,"value":870}," (default 400) — set explicitly as a cost ceiling",{"type":42,"tag":411,"props":872,"children":873},{},[874,879],{"type":42,"tag":60,"props":875,"children":876},{},[877],{"type":48,"value":878},"Shape",{"type":48,"value":880},": Reduce MinExecutionEnvironments to lower capacity during off-hours (minimum 3 for AZ resiliency)",{"type":42,"tag":452,"props":882,"children":884},{"id":883},"instance-sizing",[885],{"type":48,"value":886},"Instance Sizing",{"type":42,"tag":407,"props":888,"children":889},{},[890,900],{"type":42,"tag":411,"props":891,"children":892},{},[893,898],{"type":42,"tag":60,"props":894,"children":895},{},[896],{"type":48,"value":897},"1 vCPU + 1 GB reserved per instance",{"type":48,"value":899}," for OS overhead (not available to your function)",{"type":42,"tag":411,"props":901,"children":902},{},[903],{"type":48,"value":904},"Usable capacity = total - overhead",{"type":42,"tag":452,"props":906,"children":908},{"id":907},"configuration",[909],{"type":48,"value":910},"Configuration",{"type":42,"tag":407,"props":912,"children":913},{},[914,919,924,929,934],{"type":42,"tag":411,"props":915,"children":916},{},[917],{"type":48,"value":918},"Start with 4:1 ratio and runtime default concurrency",{"type":42,"tag":411,"props":920,"children":921},{},[922],{"type":48,"value":923},"Use ARM (Graviton) unless x86 dependencies exist",{"type":42,"tag":411,"props":925,"children":926},{},[927],{"type":48,"value":928},"Let Lambda choose instance types unless specific hardware needed",{"type":42,"tag":411,"props":930,"children":931},{},[932],{"type":48,"value":933},"Set MaxVCpuCount to control cost ceiling",{"type":42,"tag":411,"props":935,"children":936},{},[937],{"type":48,"value":938},"Never set MinExecutionEnvironments below 3 (breaks AZ resiliency)",{"type":42,"tag":452,"props":940,"children":942},{"id":941},"migration",[943],{"type":48,"value":944},"Migration",{"type":42,"tag":407,"props":946,"children":947},{},[948,953,958,963,968],{"type":42,"tag":411,"props":949,"children":950},{},[951],{"type":48,"value":952},"Start with I\u002FO-heavy functions (benefit most from multi-concurrency)",{"type":42,"tag":411,"props":954,"children":955},{},[956],{"type":48,"value":957},"Review code for concurrency safety before attaching to capacity provider",{"type":42,"tag":411,"props":959,"children":960},{},[961],{"type":48,"value":962},"Use weighted aliases for gradual traffic shift",{"type":42,"tag":411,"props":964,"children":965},{},[966],{"type":48,"value":967},"Include request IDs in all log statements",{"type":42,"tag":411,"props":969,"children":970},{},[971],{"type":48,"value":972},"Initialize DB pools and SDK clients outside the handler",{"type":42,"tag":452,"props":974,"children":976},{"id":975},"operations",[977],{"type":48,"value":978},"Operations",{"type":42,"tag":407,"props":980,"children":981},{},[982,987,992,997],{"type":42,"tag":411,"props":983,"children":984},{},[985],{"type":48,"value":986},"Set CloudWatch alarms on throttle rate > 1% and CPU > 80%",{"type":42,"tag":411,"props":988,"children":989},{},[990],{"type":48,"value":991},"Plan for 14-day instance rotation (automatic)",{"type":42,"tag":411,"props":993,"children":994},{},[995],{"type":48,"value":996},"Never manually terminate LMI EC2 instances (delete the capacity provider instead)",{"type":42,"tag":411,"props":998,"children":999},{},[1000],{"type":48,"value":1001},"Always publish a version — unpublished functions cannot run on LMI",{"type":42,"tag":89,"props":1003,"children":1005},{"id":1004},"limits-quick-reference",[1006],{"type":48,"value":1007},"Limits Quick Reference",{"type":42,"tag":96,"props":1009,"children":1010},{},[1011,1027],{"type":42,"tag":100,"props":1012,"children":1013},{},[1014],{"type":42,"tag":104,"props":1015,"children":1016},{},[1017,1022],{"type":42,"tag":108,"props":1018,"children":1019},{},[1020],{"type":48,"value":1021},"Resource",{"type":42,"tag":108,"props":1023,"children":1024},{},[1025],{"type":48,"value":1026},"Limit",{"type":42,"tag":124,"props":1028,"children":1029},{},[1030,1043,1056,1069,1082,1095,1107],{"type":42,"tag":104,"props":1031,"children":1032},{},[1033,1038],{"type":42,"tag":131,"props":1034,"children":1035},{},[1036],{"type":48,"value":1037},"Memory",{"type":42,"tag":131,"props":1039,"children":1040},{},[1041],{"type":48,"value":1042},"2 GB min, 32 GB max",{"type":42,"tag":104,"props":1044,"children":1045},{},[1046,1051],{"type":42,"tag":131,"props":1047,"children":1048},{},[1049],{"type":48,"value":1050},"Execution environments",{"type":42,"tag":131,"props":1052,"children":1053},{},[1054],{"type":48,"value":1055},"3 minimum (MinExecutionEnvironments, AZ resiliency)",{"type":42,"tag":104,"props":1057,"children":1058},{},[1059,1064],{"type":42,"tag":131,"props":1060,"children":1061},{},[1062],{"type":48,"value":1063},"Instance lifespan",{"type":42,"tag":131,"props":1065,"children":1066},{},[1067],{"type":48,"value":1068},"14 days (auto-replaced)",{"type":42,"tag":104,"props":1070,"children":1071},{},[1072,1077],{"type":42,"tag":131,"props":1073,"children":1074},{},[1075],{"type":48,"value":1076},"Concurrency\u002FvCPU",{"type":42,"tag":131,"props":1078,"children":1079},{},[1080],{"type":48,"value":1081},"64 (Node.js), 32 (Java\u002F.NET), 16 (Python)",{"type":42,"tag":104,"props":1083,"children":1084},{},[1085,1090],{"type":42,"tag":131,"props":1086,"children":1087},{},[1088],{"type":48,"value":1089},"Runtimes",{"type":42,"tag":131,"props":1091,"children":1092},{},[1093],{"type":48,"value":1094},"Node.js 22+, Java 21+, .NET 8+, Python 3.13+, Rust (provided.al2023)",{"type":42,"tag":104,"props":1096,"children":1097},{},[1098,1102],{"type":42,"tag":131,"props":1099,"children":1100},{},[1101],{"type":48,"value":592},{"type":42,"tag":131,"props":1103,"children":1104},{},[1105],{"type":48,"value":1106},"C, M, R (.large and up)",{"type":42,"tag":104,"props":1108,"children":1109},{},[1110,1114],{"type":42,"tag":131,"props":1111,"children":1112},{},[1113],{"type":48,"value":658},{"type":42,"tag":131,"props":1115,"children":1116},{},[1117],{"type":48,"value":1118},"Burst headroom equals unused capacity from TargetResourceUtilization; new instances launch within minutes",{"type":42,"tag":89,"props":1120,"children":1122},{"id":1121},"security-considerations",[1123],{"type":48,"value":1124},"Security Considerations",{"type":42,"tag":407,"props":1126,"children":1127},{},[1128,1154,1164,1174,1184,1194,1204],{"type":42,"tag":411,"props":1129,"children":1130},{},[1131,1136,1138,1144,1146,1152],{"type":42,"tag":60,"props":1132,"children":1133},{},[1134],{"type":48,"value":1135},"Operator role scoping",{"type":48,"value":1137},": Add ",{"type":42,"tag":528,"props":1139,"children":1141},{"className":1140},[],[1142],{"type":48,"value":1143},"aws:SourceAccount",{"type":48,"value":1145}," and ",{"type":42,"tag":528,"props":1147,"children":1149},{"className":1148},[],[1150],{"type":48,"value":1151},"aws:SourceArn",{"type":48,"value":1153}," conditions to trust policies to prevent confused deputy attacks.",{"type":42,"tag":411,"props":1155,"children":1156},{},[1157,1162],{"type":42,"tag":60,"props":1158,"children":1159},{},[1160],{"type":48,"value":1161},"VPC egress",{"type":48,"value":1163},": Scope security group egress to VPC endpoint security groups or AWS prefix lists rather than 0.0.0.0\u002F0.",{"type":42,"tag":411,"props":1165,"children":1166},{},[1167,1172],{"type":42,"tag":60,"props":1168,"children":1169},{},[1170],{"type":48,"value":1171},"Credentials",{"type":48,"value":1173},": Use AWS Secrets Manager or Parameter Store for database credentials — never environment variables for secrets.",{"type":42,"tag":411,"props":1175,"children":1176},{},[1177,1182],{"type":42,"tag":60,"props":1178,"children":1179},{},[1180],{"type":48,"value":1181},"Encryption",{"type":48,"value":1183},": Enable SQS SSE, CloudWatch Logs encryption (KMS), and S3 default encryption for any data at rest.",{"type":42,"tag":411,"props":1185,"children":1186},{},[1187,1192],{"type":42,"tag":60,"props":1188,"children":1189},{},[1190],{"type":48,"value":1191},"Logging",{"type":48,"value":1193},": Set CloudWatch Log group retention policies. Avoid logging PII or credentials. Enable CloudTrail data events for Lambda.",{"type":42,"tag":411,"props":1195,"children":1196},{},[1197,1202],{"type":42,"tag":60,"props":1198,"children":1199},{},[1200],{"type":48,"value":1201},"Instance rotation",{"type":48,"value":1203},": The 14-day automatic rotation ensures security patches are applied without manual intervention.",{"type":42,"tag":411,"props":1205,"children":1206},{},[1207,1212,1214,1221,1223],{"type":42,"tag":60,"props":1208,"children":1209},{},[1210],{"type":48,"value":1211},"References",{"type":48,"value":1213},": ",{"type":42,"tag":68,"props":1215,"children":1218},{"href":1216,"rel":1217},"https:\u002F\u002Fdocs.aws.amazon.com\u002Flambda\u002Flatest\u002Fdg\u002Flambda-security.html",[72],[1219],{"type":48,"value":1220},"Lambda Security Best Practices",{"type":48,"value":1222},", ",{"type":42,"tag":68,"props":1224,"children":1227},{"href":1225,"rel":1226},"https:\u002F\u002Fdocs.aws.amazon.com\u002FIAM\u002Flatest\u002FUserGuide\u002Fbest-practices.html",[72],[1228],{"type":48,"value":1229},"IAM Best Practices",{"type":42,"tag":89,"props":1231,"children":1233},{"id":1232},"files",[1234],{"type":48,"value":1235},"Files",{"type":42,"tag":96,"props":1237,"children":1238},{},[1239,1255],{"type":42,"tag":100,"props":1240,"children":1241},{},[1242],{"type":42,"tag":104,"props":1243,"children":1244},{},[1245,1250],{"type":42,"tag":108,"props":1246,"children":1247},{},[1248],{"type":48,"value":1249},"File",{"type":42,"tag":108,"props":1251,"children":1252},{},[1253],{"type":48,"value":1254},"Content",{"type":42,"tag":124,"props":1256,"children":1257},{},[1258,1273,1288,1303,1318,1333],{"type":42,"tag":104,"props":1259,"children":1260},{},[1261,1268],{"type":42,"tag":131,"props":1262,"children":1263},{},[1264],{"type":42,"tag":68,"props":1265,"children":1266},{"href":302},[1267],{"type":48,"value":305},{"type":42,"tag":131,"props":1269,"children":1270},{},[1271],{"type":48,"value":1272},"Pricing analysis, break-even calculations, Savings Plans\u002FRI impact",{"type":42,"tag":104,"props":1274,"children":1275},{},[1276,1283],{"type":42,"tag":131,"props":1277,"children":1278},{},[1279],{"type":42,"tag":68,"props":1280,"children":1281},{"href":320},[1282],{"type":48,"value":323},{"type":42,"tag":131,"props":1284,"children":1285},{},[1286],{"type":48,"value":1287},"Instance selection, memory ratios, scaling tuning, capacity provider config",{"type":42,"tag":104,"props":1289,"children":1290},{},[1291,1298],{"type":42,"tag":131,"props":1292,"children":1293},{},[1294],{"type":42,"tag":68,"props":1295,"children":1296},{"href":338},[1297],{"type":48,"value":341},{"type":42,"tag":131,"props":1299,"children":1300},{},[1301],{"type":48,"value":1302},"Concurrency model per runtime, code review checklist, Powertools compatibility",{"type":42,"tag":104,"props":1304,"children":1305},{},[1306,1313],{"type":42,"tag":131,"props":1307,"children":1308},{},[1309],{"type":42,"tag":68,"props":1310,"children":1311},{"href":356},[1312],{"type":48,"value":359},{"type":42,"tag":131,"props":1314,"children":1315},{},[1316],{"type":48,"value":1317},"Before\u002Fafter code by runtime, connection pooling, gradual cutover",{"type":42,"tag":104,"props":1319,"children":1320},{},[1321,1328],{"type":42,"tag":131,"props":1322,"children":1323},{},[1324],{"type":42,"tag":68,"props":1325,"children":1326},{"href":374},[1327],{"type":48,"value":377},{"type":42,"tag":131,"props":1329,"children":1330},{},[1331],{"type":48,"value":1332},"IAM roles, VPC setup, SAM templates, CLI commands",{"type":42,"tag":104,"props":1334,"children":1335},{},[1336,1343],{"type":42,"tag":131,"props":1337,"children":1338},{},[1339],{"type":42,"tag":68,"props":1340,"children":1341},{"href":392},[1342],{"type":48,"value":395},{"type":42,"tag":131,"props":1344,"children":1345},{},[1346],{"type":48,"value":1347},"Common errors, throttling, debugging, stuck deployments",{"items":1349,"total":1449},[1350,1367,1382,1397,1410,1420,1433],{"slug":1351,"name":1351,"fn":1352,"description":1353,"org":1354,"tags":1355,"stars":24,"repoUrl":25,"updatedAt":1366},"agents-build","add capabilities to existing agent projects","Use when adding capabilities to an existing agent project — memory, app integration, VPC, multi-agent, migration, model changes, browser, code interpreter, or resource removal. Triggers on: \"add memory\", \"remember across sessions\", \"call agent from app\", \"invoke agent from code\", \"auth to call agent\", \"streaming responses\", \"VPC\", \"VPC connectivity\", \"VPC error\", \"can't reach from VPC\", \"multi-agent\", \"A2A\", \"A2A auth\", \"orchestrator not delegating\", \"specialist not called\", \"migrate Bedrock Agent\", \"after import\", \"migration issue\", \"framework for migration\", \"change model\", \"browser tool\", \"code interpreter\", \"delete agent\", \"tear down\", \"agentcore remove\", \"cross-account memory\", \"resource-based policy on memory\", \"pay for x402 content\", \"402 Payment Required\", \"microtransactions\", \"paid API or tool\". Not for connecting to external APIs via Gateway — use agents-connect. Not for scaffolding a new project — use agents-get-started. Not for CLI\u002Fdev server errors — use agents-debug. Strands vs LangGraph in a migration context routes here.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1356,1359,1362,1363],{"name":1357,"slug":1358,"type":15},"Agents","agents",{"name":1360,"slug":1361,"type":15},"Automation","automation",{"name":23,"slug":8,"type":15},{"name":1364,"slug":1365,"type":15},"Engineering","engineering","2026-07-12T08:42:53.812877",{"slug":1368,"name":1368,"fn":1369,"description":1370,"org":1371,"tags":1372,"stars":24,"repoUrl":25,"updatedAt":1381},"agents-connect","connect agents to external services","Use when connecting your agent to external APIs, tools, or services via Gateway, or restricting tool access with Cedar policies. Handles gateway setup, target types, outbound auth (OAuth, API key, IAM), credentials, and Cedar policy authoring. Triggers on: \"connect to API\", \"add gateway\", \"connect to MCP server\", \"Lambda tools\", \"OpenAPI\", \"gateway target\", \"Cedar policy\", \"restrict tools\", \"policy engine\", \"gateway auth error\", \"store API key\", \"outbound credential\", \"env var API key\", \"API key None after deploy\", \"credential not available after deploy\", \"should this be a gateway target\", \"give my agent tools\", \"add tools to agent\". Not for inbound auth (who can call your agent) — use agents-harden. Not for debugging agent behavior — use agents-debug. Not for VPC networking errors (agent can't reach APIs due to VPC) — use agents-build. Not for creating or hosting a new MCP server project — use agents-get-started.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1373,1374,1377,1380],{"name":1357,"slug":1358,"type":15},{"name":1375,"slug":1376,"type":15},"API Development","api-development",{"name":1378,"slug":1379,"type":15},"Authentication","authentication",{"name":23,"slug":8,"type":15},"2026-07-16T06:00:38.866147",{"slug":1383,"name":1383,"fn":1384,"description":1385,"org":1386,"tags":1387,"stars":24,"repoUrl":25,"updatedAt":1396},"agents-debug","debug agent and environment issues","Use when your agent or environment is broken — wrong answers, errors, timeouts, tool failures, or CLI issues. Reads traces and logs to diagnose root causes. Also checks prerequisites when the CLI itself isn't working. Triggers on: \"agent not working\", \"wrong answer\", \"agent error\", \"tool call failing\", \"debug agent\", \"check logs\", \"read traces\", \"broken\", \"500 error\", \"424 error\", \"model access denied\", \"command not found\", \"stuck in DELETING\", \"maxVms exceeded\", \"cold start diagnosis\", \"cold start slow\", \"agentcore create error\", \"create failed\", \"exit code 7\", \"connection refused local dev\". Not for deploy failures — use agents-deploy. Not for performance tuning without errors — use agents-optimize. Not for VPC configuration — use agents-build. Not for observability setup or missing logs — use agents-optimize.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1388,1389,1390,1393],{"name":1357,"slug":1358,"type":15},{"name":23,"slug":8,"type":15},{"name":1391,"slug":1392,"type":15},"Debugging","debugging",{"name":1394,"slug":1395,"type":15},"Observability","observability","2026-07-16T06:00:44.679093",{"slug":1398,"name":1398,"fn":1399,"description":1400,"org":1401,"tags":1402,"stars":24,"repoUrl":25,"updatedAt":1409},"agents-deploy","deploy AI agents to AWS","Use when deploying your agent to AWS, or when a deploy has failed. Handles pre-flight validation, CDK\u002FIAM\u002Fquota error diagnosis, version management, rollback, and canary deployments. Triggers on: \"deploy my agent\", \"agentcore deploy\", \"deploy failed\", \"CDK error\", \"rollback\", \"canary deploy\", \"pin version\", \"redeploy\", \"deploy stuck\". Not for production hardening — use agents-harden. Not for adding capabilities before deploy — use agents-build or agents-connect. Not for VPC configuration errors — use agents-build.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1403,1404,1405,1408],{"name":1357,"slug":1358,"type":15},{"name":23,"slug":8,"type":15},{"name":1406,"slug":1407,"type":15},"CI\u002FCD","ci-cd",{"name":17,"slug":18,"type":15},"2026-07-12T08:42:55.059577",{"slug":1411,"name":1411,"fn":1412,"description":1413,"org":1414,"tags":1415,"stars":24,"repoUrl":25,"updatedAt":1419},"agents-get-started","scaffold and deploy new agent projects","Use when a developer wants to create a new agent project or get started with AgentCore. Handles framework selection, project scaffolding, first deploy, and first invocation. Triggers on: \"build an agent\", \"create an agent\", \"get started\", \"new project\", \"agentcore create\", \"which framework\", \"Strands vs LangGraph\", \"hello world agent\", \"first agent\", \"create MCP server\", \"host MCP server\", \"agentcore dev\", \"dev server\", \"what port\", \"local development\". Not for adding capabilities to existing projects — use agents-build or agents-connect. Strands vs LangGraph in a migration context routes to agents-build, not here. Connecting to an existing MCP server routes to agents-connect, not here.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1416,1417,1418],{"name":1357,"slug":1358,"type":15},{"name":23,"slug":8,"type":15},{"name":17,"slug":18,"type":15},"2026-07-12T08:42:51.963247",{"slug":1421,"name":1421,"fn":1422,"description":1423,"org":1424,"tags":1425,"stars":24,"repoUrl":25,"updatedAt":1432},"agents-harden","harden agents for production","Use when preparing your agent for production — IAM scoping, inbound auth (JWT, SigV4), secrets management, cold start optimization, session lifecycle, rate limiting, input validation, and quota guidance. Triggers on: \"production checklist\", \"harden agent\", \"production ready\", \"secure agent\", \"inbound auth\", \"going live\", \"cold start optimization\", \"session lifecycle\", \"StopRuntimeSession\", \"quota\", \"throttling\", \"maxVms\", \"rate limit\", \"security audit of outbound API calls\", \"gateway target audit for production\", \"restrict who can call\", \"lock down endpoint\", \"only our app can call\". Not for Cedar tool-restriction policies — use agents-connect. Not for quality measurement — use agents-optimize. Not for outbound credential storage or API key wiring — use agents-connect. Not for A2A agent-to-agent auth — use agents-build. Cold start observation and diagnosis (not optimization) routes to agents-debug.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1426,1427,1428,1429],{"name":1357,"slug":1358,"type":15},{"name":23,"slug":8,"type":15},{"name":785,"slug":782,"type":15},{"name":1430,"slug":1431,"type":15},"Security","security","2026-07-16T06:00:42.174705",{"slug":1434,"name":1434,"fn":1435,"description":1436,"org":1437,"tags":1438,"stars":24,"repoUrl":25,"updatedAt":1448},"agents-optimize","optimize agent quality and performance","Use when measuring or improving agent quality and performance — set up evaluators, online monitoring, CI\u002FCD quality gates, observability, or cost optimization. Triggers on: \"evaluate my agent\", \"add evaluator\", \"measure quality\", \"quality gate\", \"run evals\", \"agent too slow\", \"why is it slow\", \"reduce latency\", \"set up observability\", \"CloudWatch dashboard\", \"how much does my agent cost\", \"cost optimization\", \"logs not showing up\", \"logs missing\", \"spans not found\", \"eval failing\", \"eval error\", \"dev traces\", \"local traces\", \"agentcore dev traces\", \"traces to CloudWatch\". Not for debugging errors or crashes — use agents-debug. Slow but correct routes here; broken routes to debug.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1439,1440,1441,1444,1445],{"name":1357,"slug":1358,"type":15},{"name":23,"slug":8,"type":15},{"name":1442,"slug":1443,"type":15},"Evals","evals",{"name":1394,"slug":1395,"type":15},{"name":1446,"slug":1447,"type":15},"Performance","performance","2026-07-12T08:42:56.488105",114,{"items":1451,"total":1568},[1452,1459,1466,1473,1480,1486,1493,1501,1518,1531,1543,1558],{"slug":1351,"name":1351,"fn":1352,"description":1353,"org":1453,"tags":1454,"stars":24,"repoUrl":25,"updatedAt":1366},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1455,1456,1457,1458],{"name":1357,"slug":1358,"type":15},{"name":1360,"slug":1361,"type":15},{"name":23,"slug":8,"type":15},{"name":1364,"slug":1365,"type":15},{"slug":1368,"name":1368,"fn":1369,"description":1370,"org":1460,"tags":1461,"stars":24,"repoUrl":25,"updatedAt":1381},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1462,1463,1464,1465],{"name":1357,"slug":1358,"type":15},{"name":1375,"slug":1376,"type":15},{"name":1378,"slug":1379,"type":15},{"name":23,"slug":8,"type":15},{"slug":1383,"name":1383,"fn":1384,"description":1385,"org":1467,"tags":1468,"stars":24,"repoUrl":25,"updatedAt":1396},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1469,1470,1471,1472],{"name":1357,"slug":1358,"type":15},{"name":23,"slug":8,"type":15},{"name":1391,"slug":1392,"type":15},{"name":1394,"slug":1395,"type":15},{"slug":1398,"name":1398,"fn":1399,"description":1400,"org":1474,"tags":1475,"stars":24,"repoUrl":25,"updatedAt":1409},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1476,1477,1478,1479],{"name":1357,"slug":1358,"type":15},{"name":23,"slug":8,"type":15},{"name":1406,"slug":1407,"type":15},{"name":17,"slug":18,"type":15},{"slug":1411,"name":1411,"fn":1412,"description":1413,"org":1481,"tags":1482,"stars":24,"repoUrl":25,"updatedAt":1419},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1483,1484,1485],{"name":1357,"slug":1358,"type":15},{"name":23,"slug":8,"type":15},{"name":17,"slug":18,"type":15},{"slug":1421,"name":1421,"fn":1422,"description":1423,"org":1487,"tags":1488,"stars":24,"repoUrl":25,"updatedAt":1432},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1489,1490,1491,1492],{"name":1357,"slug":1358,"type":15},{"name":23,"slug":8,"type":15},{"name":785,"slug":782,"type":15},{"name":1430,"slug":1431,"type":15},{"slug":1434,"name":1434,"fn":1435,"description":1436,"org":1494,"tags":1495,"stars":24,"repoUrl":25,"updatedAt":1448},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1496,1497,1498,1499,1500],{"name":1357,"slug":1358,"type":15},{"name":23,"slug":8,"type":15},{"name":1442,"slug":1443,"type":15},{"name":1394,"slug":1395,"type":15},{"name":1446,"slug":1447,"type":15},{"slug":1502,"name":1502,"fn":1503,"description":1504,"org":1505,"tags":1506,"stars":24,"repoUrl":25,"updatedAt":1517},"amazon-aurora-mysql","manage Amazon Aurora MySQL clusters","Amazon Aurora MySQL — creates, modifies, and advises on Aurora MySQL clusters specifically (MySQL-compatible engine, Aurora serverless, parallel query). Trigger for Aurora MySQL cluster operations, ACU sizing, I\u002FO-Optimized storage, commitment pricing, or MySQL upgrade planning. Aurora MySQL uses full (VPC-based) configuration — express configuration is PostgreSQL-only. For Aurora PostgreSQL, use amazon-aurora-postgresql instead. Contains safety guardrails and response templates that override defaults.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1507,1508,1511,1514],{"name":23,"slug":8,"type":15},{"name":1509,"slug":1510,"type":15},"Database","database",{"name":1512,"slug":1513,"type":15},"MySQL","mysql",{"name":1515,"slug":1516,"type":15},"Serverless","serverless","2026-07-12T08:43:13.27939",{"slug":1519,"name":1519,"fn":1520,"description":1521,"org":1522,"tags":1523,"stars":24,"repoUrl":25,"updatedAt":1530},"amazon-aurora-postgresql","configure Amazon Aurora PostgreSQL clusters","Amazon Aurora PostgreSQL — creates, modifies, and advises on Aurora PostgreSQL clusters specifically (PostgreSQL-compatible engine, Aurora serverless, express configuration, pgvector, Babelfish). Trigger for Aurora PostgreSQL cluster operations, express-configuration quick-start, ACU sizing, I\u002FO-Optimized storage, commitment pricing, or PostgreSQL upgrade planning. For Aurora MySQL, use amazon-aurora-mysql instead. Contains safety guardrails, express-first routing, and response templates that override defaults.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1524,1525,1526,1529],{"name":23,"slug":8,"type":15},{"name":1509,"slug":1510,"type":15},{"name":1527,"slug":1528,"type":15},"PostgreSQL","postgresql",{"name":1515,"slug":1516,"type":15},"2026-07-16T06:00:34.789624",{"slug":1532,"name":1532,"fn":1533,"description":1534,"org":1535,"tags":1536,"stars":24,"repoUrl":25,"updatedAt":1542},"amazon-bedrock","build generative AI apps with Amazon Bedrock","Builds generative AI applications on Amazon Bedrock. Covers model invocation (Converse API, InvokeModel), RAG with Knowledge Bases, Bedrock Agents, Guardrails, and AgentCore. Use when invoking models, setting up Knowledge Bases, creating agents, applying guardrails, deploying to AgentCore, migrating\u002Fporting\u002Fconverting a Bedrock Agent (including inline agents) to an AgentCore Harness, troubleshooting Bedrock errors (ThrottlingException, AccessDeniedException), or choosing models (Claude, Llama, Nova, Titan). ALSO USE for prompt caching setup and debugging, quota health checks and throttling diagnosis, cost attribution and tracking, migrating between Claude model generations (4.5 to 4.6 to 4.7), chunking strategies, API selection (Converse vs InvokeModel), guardrail capabilities, and model selection. Also covers AgentCore Payments setup (x402, microtransactions, Payment Manager, Connector, Instrument, Coinbase CDP, Stripe Privy, 402 Payment Required, pay for content, paid endpoint, agent payments). NOT for custom model training, Rekognition, or Comprehend.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1537,1538,1539],{"name":1357,"slug":1358,"type":15},{"name":23,"slug":8,"type":15},{"name":1540,"slug":1541,"type":15},"LLM","llm","2026-07-25T05:30:35.20899",{"slug":1544,"name":1544,"fn":1545,"description":1546,"org":1547,"tags":1548,"stars":24,"repoUrl":25,"updatedAt":1557},"amazon-documentdb","manage Amazon DocumentDB clusters","Manages Amazon DocumentDB end-to-end — serverless-on-8.0 cluster setup, TLS\u002FVPC\u002Fdriver config, flexible-schema and vector-search data modeling, MongoDB compatibility assessment, DMS-based migration, slow-query diagnosis, major version upgrades (4.0→5.0→8.0), Well-Architected reviews (41-check wa_review.py), cost estimation, and security hardening. Retrieve for every DocumentDB question and when the user asks to set up or migrate MongoDB to AWS — DocumentDB is AWS's MongoDB-compatible managed database. Triggers: JSON document store, document database, MongoDB on AWS, Nested fields, Lambda cannot connect, TLS handshake, VPC port 27017, IAM auth, Secrets Manager, encryption at rest, $graphLookup, flexible schema, COLLSCAN, compound index, DMS migration, CDC cutover, $vectorSearch, RAG, Global Clusters, DR replication, cost sizing, audit, health check, production-readiness.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1549,1550,1551,1554],{"name":23,"slug":8,"type":15},{"name":1509,"slug":1510,"type":15},{"name":1552,"slug":1553,"type":15},"MongoDB","mongodb",{"name":1555,"slug":1556,"type":15},"NoSQL","nosql","2026-07-12T08:43:00.455878",{"slug":1559,"name":1559,"fn":1560,"description":1561,"org":1562,"tags":1563,"stars":24,"repoUrl":25,"updatedAt":1567},"amazon-dynamodb","design and debug DynamoDB data layers","Designs, reviews, and debugs DynamoDB data layers from design axioms — enumerates access patterns, chooses partition\u002Fsort keys and GSIs, decides single-table vs. multi-table, configures Streams, Global Tables, TTL, and zero-ETL integrations to OpenSearch\u002FRedshift\u002FSageMaker Lakehouse, and produces a defensible data-layer design with a monthly cost estimate and optional live validation. Applies whenever a user is designing, reviewing, or refactoring anything backed by DynamoDB — schemas, access patterns, GSIs, single- vs. multi-table choices, Streams consumers, transactional outboxes, Global Tables, zero-ETL pipelines — even when they don't say \"axioms\" or \"design review.\" Also applies when debugging hot partitions, throttling, unbounded Scans, LWW conflicts, or surprise bills on DynamoDB workloads.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1564,1565,1566],{"name":23,"slug":8,"type":15},{"name":1509,"slug":1510,"type":15},{"name":1555,"slug":1556,"type":15},"2026-07-16T06:00:37.690386",115]