Tag Configuration

Learn how to apply settings via cloud provider tags.

Sedai automatically infers resource tags and Kubernetes annotations. These can be used to optionally define feature settings and compute action controls as well as for configuring Infrastructure as Code (IaC) changes.

---

Features

When you define settings via tags, Sedai considers it an override on settings a resource would otherwise inherit from its parent (such as a group, account, or Kubernetes cluster).

Tags must include the prefix settings.sedai.io followed by a period (.); for example:

settings.sedai.io.optimization.setting.configMode

For feature settings, the accepted keys of MANUAL and AUTO correspond to Recommend Mode and Autonomous Mode respectively.

For Kubernetes workloads, add the prefix settings.sedai.io followed by a slash (/); for example:

settings.sedai.io/optimization.setting.configMode

For feature settings, the accepted keys of MANUAL and AUTO correspond to Copilot mode and Autopilot mode respectively.

Optimization

Learn more about optimization settings.

AWS ECS/Kubernetes

Setting	Key	Accepted Key
Optimization Setting	optimization.setting.configMode	OFF, MANUAL, AUTO
Goal	optimization.optimizationFocus.focus	COST, DURATION, COST_AND_DURATION
If goal is to improve performance, the allowed percent increase impact on cost from memory	optimization.optimizationFocus.maxMemoryIncreasePct	Integer between 0-100
If goal is to improve performance, the allowed percent increase impact on cost from CPU	optimization.optimizationFocus.maxCPUIncreasePct	Integer between 0-100
If goal is to reduce cost, the allowed percent increase on latency	optimization.optimizationFocus.maxLatencyIncreasePct	Integer between 0-100

AWS Lambda

Setting	Key	Accepted Key
Memory optimization setting	optimization.setting.configMode	OFF, MANUAL, AUTO
Goal	optimization.optimizationFocus.focus	COST, DURATION, COST_AND_DURATION
If goal is to improve performance, the allowed percent increase impact on cost	optimization.optimizationFocus.autoOptMaxCostChangePct	Integer between 0-100
If goal is to reduce cost, the allowed percent increase on duration	optimization.optimizationFocus.autoOptMaxLatencyChangePct	Integer between 0-100
Autonomous Concurrency setting	optimization.concurrency.configMode	OFF, MANUAL, AUTO

Availability

Learn more about availability settings.

Key	Acceoted Value
availability.configMode	OFF, MANUAL, AUTO
AWS Lambda only telemetryLogging.enabled	Boolean

Release Intelligence

Learn more about release intelligence settings.

Key	Acceoted Value
releaseIntelligence.configMode	OFF, MANUAL

Compute Actions

The following settings only apply to AWS ECS/Fargate and Kubernetes stateless workloads.

Learn more about container and virtual machine compute actions.

Setting	Key	Accepted Value
Vertical Scaling	enableVerticalScaling.enabled	Boolean
↳ Minimum CPU in MiB (optional)	enableVerticalScaling.minCpu	Integer
↳ Minimum memory in GB (optional)	enableVerticalScaling.minMemory	Integer
Horizontal Scaling	enableHorizontalScaling.enabled	Boolean
↳ Minimum replica count (optional) Default is set to 2	enableHorizontalScaling.minReplicas	Integer
↳ Maximum replica count (optional)	enableHorizontalScaling.maxReplicas	Integer
↳ ECS only: Replica increment count (optional)	enableHorizontalScaling.replicaMultiplier	Integer
↳ ECS only: Replica increment count (optional)	enableHorizontalScaling.replicaIncrement	Integer
Auto Scaling (ECS only)	enableServiceAutoscalingConfiguration.enabled	Boolean
Autonomous Action without Traffic	autonomousActionWithoutTraffic.enabled	Boolean
Pre-production/Production	isProd.enabled	Boolean

---

IaC Configurations

Requires an IaC integration — learn more.

To apply IaC configurations, tags should be formatted with the following prefix:

configs:sedai.io:

For Kubernetes workloads, annotations should be formatted with the prefix:

configs.sedai.io/

For example:

Tag	Example
default_repo_path	configs:sedai.io:default_repo_path = 412335
variables_file_path	configs:sedai.io:variables_file_path = terraform/prod/prod-sls-1.tfvars

If you use GitLabs, the variables_file_path does not need to include the project name.

For example, if the project name is A and the variables file C is at location A/B/C then you only need to include B/C as the value for variables_file_path.

---

AWS Lambda

Tag	Example
memory_size (MB)	configs:sedai:io:memory_size = var.memory_size["prod-sls-1"]
timeout (seconds)	configs:sedai.io:timeout = var.timeout[“prod-sls-1”]
reserved_concurrency	configs:sedai.io:reserved_concurrency = var.reserved_concurrency[“prod-sls-1”]
provisioned_concurrency	configs:sedai.io:provisioned_concurrency = var.provisioned_concurrency[“prod-sls-1”]

AWS ECS

Tag	Example
task_cpu (units)	configs:sedai.io:task_cpu = var.task_cpu[“prod-app-1”]
task_memory (MiB)	configs:sedai.io:task_memory = var.task_memory[“prod-app-1”]
desired_count	configs:sedai.io:desired_count = var.desired_count[“prod-app-1”]
container.soft_memory (MiB)	configs:sedai.io:container.soft_memory = var.soft_memory[“prod-app-1”]
container.hard_memory (MiB)	configs:sedai.io:container.hard_memory = var.hard_memory[“prod-app-1”]
container.soft_cpu (units)	configs:sedai.io:container.soft_cpu = var.memosoft_cpury_size[“prod-app-1”]
autoscaler_config.as_min_task	configs:sedai.io:autoscaler_config.as_min_task = var.as_min_task[“prod-app-1”]
autoscaler_config.as_max_task	configs:sedai.io:autoscaler_config.as_max_task = var.as_max_task[“prod-app-1”]
autoscaler_config.metric	configs:sedai.io:autoscaler_config.metric = var.metric[“prod-app-1”]
autoscaler_config.target_value	configs:sedai.io:autoscaler_config.target_value = var.target_value[“prod-app-1”]

Kubernetes

Annotation	Example
replica_count	configs.sedai.io/container.replica_count = var.replica_count[“prod-app-1”]
container.cpu_request	configs.sedai.io/container.cpu_request = var.cpu_request[“prod-app-1”]
container.cpu_limit	configs.sedai.io/container.cpu_limit = var.cpu_limit[“prod-app-1”]
container.memory_request	configs.sedai.io/container.memory_request = var.memory_request[“prod-app-1”]
container.memory_limit	configs.sedai.io/container.memory_limit = var.memory_limit[“prod-app-1”]
hpa_min_count	configs.sedai.io/hpa_min_count = var.hpa_min_count[“prod-app-1”]
hpa_max_count	configs.sedai.io/hpa_max_count = var.hpa_max_count[“prod-app-1”]
hpa_metric	configs.sedai.io/hpa_metric = var.hpa_metric[“prod-app-1”]
hpa_target_value	configs.sedai.io/hpa_target_value = var.hpa_target_value[“prod-app-1”]