The following page lists the default limits of Davis AI components.
Some of the problems and events limits apply per provider. The following providers are subject to limits:
AGENT_LOCAL_REST_API_INGESTAVAILABILITYBASELININGEVENTS_REST_API_INGESTKUBERNETES_ANOMALY_DETECTIONKUBERNETES_EVENTLOG_EVENTSMETRIC_EVENTSONEAGENTOPENPIPELINE_DATA_EXTRACTIONREAL_USER_MONITORINGSYNTHETICThe maximum number of simultaneously active Davis problems within a single environment is 10,000.
The maximum number of simultaneously active Davis events within a single environment, across all event providers, is 15,000.
The maximum number of simultaneously active Davis events per event provider is 4,000. Example of the event provider: event.provider = "AVAILABILITY".
The maximum number of Davis events per event provider that can be processed within an hour is 100,000. This limit is replenished every hour.
Certain events override default limits. See the table below for more information.
Event provider
Maximum limit
AVAILABILITY
200,000 Davis events processed per hour.
METRIC_EVENTS
200,000 Davis events processed per hour.
AGENT_LOCAL_REST_API_INGESTBASELININGEVENTS_REST_API_INGESTKUBERNETES_ANOMALY_DETECTIONKUBERNETES_EVENTLOG_EVENTSONEAGENTOPENPIPELINE_DATA_EXTRACTIONREAL_USER_MONITORINGSYNTHETICThe default of 100,000 Davis events processed per hour applies.
The maximum lifetime for active events is 60 days. If the underlying anomaly doesn't disappear after 60 days, the respective event is closed and a new one is created.
The maximum problem merging timeframe is 90 minutes. If the problem remains unresolved for more than 90 minutes, no new events will be merged into it after that.
This measure ensures that Davis avoids collecting unrelated information for long-lasting incidents.
Item
Maximum limit
Number of anomaly detector configurations
1,000 per environment
Grail query request timeout
10 seconds
Item
Maximum limit
Description
Number of metric events
10,000 per environment
The maximum number of metric events that can be configured.
Number of monitored dimensions
100,000
This limit applies across all customer-defined metric selector metric events.
Dynatrace also counts aggregated dimensions towards the limit. Once the limit is reached, no additional dimensions will be monitored.
Number of static threshold metric event configurations
100
The maximum number of metric event configurations that use the static threshold model.
Number of static threshold monitored dimensions
1,000 per configuration
The maximum number of static threshold monitored dimensions. Once the limit is reached, no additional dimensions will be monitored.
Number of auto-adaptive threshold metric event configurations
100
The maximum number of metric event configurations that use the auto-adaptive threshold model.
Number of auto-adaptive threshold monitored dimensions
1,000 per configuration
The maximum number of auto-adaptive threshold monitored dimensions. Once the limit is reached, no additional dimensions will be monitored.
Number of seasonal baseline metric event configurations
100
The maximum number of metric event configurations that use the seasonal baseline model.
Number of seasonal baseline monitored dimensions
500 per configuration
The maximum number of seasonal baseline monitored dimensions. Once the limit is reached, no additional dimensions will be monitored.
Hard monitored dimension limit
10,000 per configuration
If the number of monitored dimensions exceeds the hard limit, the query will start failing.
Number of simultaneously active alerts
200 per configuration
The maximum number of simultaneously active alerts per metric key-based configuration.
Item
Maximum limit
Description
Data records
1,000
The maximum number of timeseries data records that can be uniquely simulated per single Davis analyzer execution.
Item
Maximum limit
Individual user requests
10 requests per 15 minutes
All user requests across the environment
40 requests per 15 minutes