CloudWatch operations
Every CloudWatch operation Overcast declares — 15 of 17 implemented — with status, behaviour notes and a link to the AWS API reference for each.
15 of 17 listed operations are implemented. Back to CloudWatch.
Summary
| Category | ✅ Supported | ❌ Unsupported |
|---|---|---|
| General | 15 | 2 |
Endpoints
General
| Operation | Status | Notes | AWS Docs |
|---|---|---|---|
DeleteAlarms | ✅ Supported | Deletes one or more alarms by name, along with their history | docs |
DescribeAlarmHistory | ✅ Supported | Returns StateUpdate, ConfigurationUpdate and Action items; filters by alarm name, item type, date range, MaxRecords and ScanBy. The most recent 100 items per alarm are retained | docs |
DescribeAlarms | ✅ Supported | Lists alarms, supports filtering; reports the evaluator’s live StateValue, StateReason and StateReasonData | docs |
DescribeAlarmsForMetric | ✅ Supported | Lists alarms for a specific metric | docs |
DisableAlarmActions | ✅ Supported | Clears ActionsEnabled so transitions stop firing actions | docs |
EnableAlarmActions | ✅ Supported | Sets ActionsEnabled so transitions fire actions again | docs |
GetMetricData | ✅ Supported | Returns query-based metric values over time ranges, over all three protocols the model declares — Query, awsJson1_0 (primary, #886) and Smithy RPC v2 CBOR (#1280) — from one shared evaluation core, so an SDK gets the same result whichever it negotiates. Merges custom PutMetricData points with automatically-recorded service metrics (AWS/Lambda’s Invocations/Errors/Duration/Throttles/ConcurrentExecutions pilot) from internal/metrics, read-through, when service-metrics collection is enabled | docs |
GetMetricStatistics | ✅ Supported | Returns aggregated datapoints by period, over Query, awsJson and rpcv2Cbor alike (#1280). Merges custom PutMetricData points with automatically-recorded service metrics from internal/metrics, so an alarm on e.g. AWS/Lambda Errors evaluates against real invocation outcomes | docs |
ListMetrics | ✅ Supported | Lists available metrics over Query, awsJson and rpcv2Cbor (#1280), merging custom PutMetricData series with automatically-recorded service metrics | docs |
ListTagsForResource | ✅ Supported | Lists tags for an alarm, over the Query, JSON and rpcv2Cbor protocols (#1280). Alarms are the only taggable CloudWatch resource Overcast emulates, so any other ResourceARN is ResourceNotFoundException — or InvalidParameterValue when it is not a CloudWatch ARN at all | docs |
PutMetricAlarm | ✅ Supported | Creates or updates a single-metric alarm, which is then evaluated automatically (Threshold, ComparisonOperator, Period, EvaluationPeriods, DatapointsToAlarm, Dimensions, TreatMissingData). Metric-math/multi-metric alarms, anomaly detection (ThresholdMetricId) and extended statistics are created but never evaluated, and say so in the alarm’s StateReason and an x-overcast-emulation-limitation response header; PromQL alarms (EvaluationCriteria) are still refused with 501. Tags are applied at creation only and validated against the same rules as TagResource, so a create carrying an invalid tag set fails rather than leaving an untagged alarm | docs |
PutMetricData | ✅ Supported | Publishes metric data points | docs |
SetAlarmState | ✅ Supported | Forces an alarm’s state and fires that state’s actions; the forced state is held against the evaluator for one evaluation range | docs |
PutAnomalyDetector | ❌ Unsupported | stub; returns 501 — there is no anomaly-detection model behind the emulator | docs |
PutCompositeAlarm | ❌ Unsupported | stub; returns 501 — composite alarm rules are not evaluated | docs |
TagResource | ✅ Supported | Adds or updates tags on an alarm, over the Query, JSON and rpcv2Cbor protocols (#1280). Tagging an alarm that does not exist is ResourceNotFoundException, as on AWS. Tag sets are validated — 50 tags per resource, keys 1-128 characters and not aws:-prefixed, values up to 256 — and a rejected set is not written | docs |
UntagResource | ✅ Supported | Removes tags from an alarm, over the Query, JSON and rpcv2Cbor protocols (#1280). A key that is not present is ignored; an alarm that does not exist is ResourceNotFoundException | docs |
Related
- CloudWatch — quick start, what works, and the differences from AWS
- All service pages