metrics) and an Integer parameter (
deviatingValue). The health state
DEVIATINGwill be returned when the last metric point in the metrics stream a metric value higher than the specified deviating value.
events). The check function will return the health state DEVIATING when the first event in the stream includes an event with value
barfor the key
Anomaly Eventswhen the check is configured.
event.typeequal to "Metric Stream No Anomaly" is emitted.
ANY(either RISE or DROP). The parameter gives fine-grained control over alerting on anomaly events. For example, the direction
RISEmay be interesting to track latency, while
DROPmight be useful to track request count.
HealthStateValue. This will be the new health state of the component (
CheckStateExpiration. Specifies how long a health state should remain valid and what it should change to after expiration.
RunStateValue. This will be the new run state of the component (
detailedMessage- Markdown formatted explanation of the reason behind a returned health state.
shortMessage- A short description of the state change.
causingEvents- The events that triggered the health state change. These are used in check functions that listen to anomaly events to link changes to anomaly events. Provided as a map with the keys
data- Arbitrary additional data.
DEVIATINGhealth status can then be triggered if an anomaly with certain parameters is found for a specified metric stream.
Metric Stream Anomalyevent and if the metric stream ID of the event matches that provided by the user for the metric stream ID parameter (
metricStream). If there is a match, the function will trigger a
DEVIATINGhealth state, hold it for 1 minute (60000 milliseconds) and then switch the health state to
UNKNOWN. Additionally, the return value
causingEventswill return details of the event that caused the state change.