specification: API Commons Rate Limits
specificationVersion: '0.1'
schema: https://raw.githubusercontent.com/api-evangelist/interface-research/main/schema/api-commons.yml#/$defs/RateLimits
provider: AWS CloudWatch
providerId: cloudwatch
created: '2026-05-04'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-04, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-05'
reconciled: true
tags:
  - Aws
  - Monitoring
  - Observability
  - Logs
  - Metrics
  - Rate Limiting
description: AWS CloudWatch enforces per-account, per-region transactions-per-second (TPS) throttling on
  every API action. Default quotas are integrated with AWS Service Quotas; most are adjustable via a quota-increase
  request. Throttled requests return HTTP 400 with a Throttling or ThrottlingException error code. Limits
  vary per service (CloudWatch core, Logs, Application Insights, Application Signals, RUM, Internet Monitor,
  Network Flow Monitor, Synthetics, OAM). Consult Service Quotas (`--service-code monitoring` etc.) for
  the live values for your account.
sources:
  - https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/cloudwatch_limits.html
  - https://docs.aws.amazon.com/general/latest/gr/cw_region.html
responseCodes:
  throttled: 400
  quotaExceeded: 400
  serviceUnavailable: 503
limits:
  - name: CloudWatch core API throttling
    scope: account/region
    metric: requests_per_second
    limit: see service-specific quotas page
    notes: Default TPS varies by API action (PutMetricData, GetMetricData, ListMetrics, PutMetricAlarm,
      etc.). Look up live values via `aws service-quotas list-service-quotas --service-code monitoring`.
  - name: CloudWatch Application Insights API throttling
    scope: account/region
    metric: requests_per_second
    limit: 5
    timeFrame: second
    notes: All Application Insights API actions are throttled to 5 TPS; this quota is not adjustable.
  - name: CloudWatch Logs API throttling
    scope: account/region
    metric: requests_per_second
    limit: see service-specific quotas page
    notes: Use `aws service-quotas list-service-quotas --service-code logs` for the live values.
  - name: CloudWatch Application Signals API throttling
    scope: account/region
    metric: requests_per_second
    limit: see service-specific quotas page
    notes: Use `--service-code application-signals`.
  - name: CloudWatch RUM API throttling
    scope: account/region
    metric: requests_per_second
    limit: see service-specific quotas page
    notes: Use `--service-code rum`.
  - name: CloudWatch Internet Monitor API throttling
    scope: account/region
    metric: requests_per_second
    limit: see service-specific quotas page
    notes: Use `--service-code internetmonitor`.
  - name: CloudWatch Network Monitor API throttling
    scope: account/region
    metric: requests_per_second
    limit: see service-specific quotas page
    notes: Use `--service-code networkmonitor`.
  - name: CloudWatch Observability Access Manager API throttling
    scope: account/region
    metric: requests_per_second
    limit: see service-specific quotas page
    notes: Use `--service-code oam`.
  - name: AI Operations API throttling
    scope: account/region
    metric: requests_per_second
    limit: see service-specific quotas page
    notes: Use `--service-code aiops`.
  - name: PutMetricData payload size
    scope: account/region
    metric: requests_per_second
    limit: see service-specific quotas page
    notes: PutMetricData accepts up to 1,000 metric data points per request and is rate-limited per account.
policies:
  - name: AWS SigV4 throttling
    description: All CloudWatch API requests are AWS SigV4 signed and throttled at the account/region level.
      A 400 response with code Throttling or ThrottlingException indicates the request was rate-limited.
  - name: Exponential backoff
    description: AWS SDKs implement exponential backoff with jitter by default. Custom clients should match
      this behavior on Throttling errors.
  - name: Service Quotas integration
    description: All CloudWatch quotas are visible in the AWS Service Quotas console (`monitoring`, `logs`,
      `application-signals`, `internetmonitor`, `networkmonitor`, `oam`, `rum`, `aiops` service codes) and
      most can be raised via a quota-increase request.
  - name: Per-region scoping
    description: Quotas are per AWS account per region; running in multiple regions effectively multiplies
      the available TPS budget.
  - name: PutMetricData batching
    description: Batch up to 1,000 metric data points per PutMetricData call to maximize throughput against
      the per-account TPS quota.
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com