APM alerts - Atatus Docs

Atatus APM alerts monitor server-side application performance to detect latency issues, error spikes, and throughput anomalies. You can alert on web transactions, background jobs, database queries, external service calls, and exceptions. APM alerts also support baseline alerts for detecting deviations from historical patterns.

Available metrics

Web transaction metrics

Metric	Summary Function	Unit
Apdex Score	apdex	apdex
Web Response Time	average	seconds
Web Throughput	throughput	calls
HTTP Failure Count	throughput	count
HTTP Failure Rate	percent	%

Background job metrics

Metric	Summary Function	Unit
Background Response Time	average	seconds
Background Throughput	throughput	calls

External service metrics

Metric	Summary Function	Unit
External Request Response Time	average	seconds
External Request Throughput	throughput	calls

Database metrics

Metric	Summary Function	Unit
Database Query Time	average	seconds
Database Throughput	throughput	calls

Error metrics

Metric	Summary Function	Unit
Exception Count	throughput	count

Enter thresholds in the unit shown. For response time metrics, enter the value in seconds (e.g., 0.5 for 500 ms, 2 for 2 seconds). The alert engine handles internal unit conversion automatically.

Rule types

APM alerts support the following rule types:

Rule Type	Description
Application Metric	Standard APM metrics (response time, throughput, errors)
External Service	External/remote service call metrics
Key Transaction	Alerts scoped to a specific key transaction
Response Time Percentile	Percentile-based response time alerts

Baseline metrics

APM alerts also support baseline metrics. See Baseline alerts for details. Supported APM baseline metrics: Web Response Time, Web Throughput, Exception Count, External Request Response Time, External Request Throughput, Database Query Time, Database Throughput.

Targets

APM alert rules are scoped by project (application). Select one or more applications as targets. Each application is evaluated independently.

How alert evaluation works

Operators

Operator	Triggers when
`above`	metric value > threshold
`below`	metric value < threshold
`equal`	metric value = threshold

Evaluation windows

Available durations: 5, 10, 15, 30, or 60 minutes.

Time functions

Function	Behavior
`all`	Triggers only if every 1-minute bucket in the window breaches the threshold.
`any`	Triggers if at least one 1-minute bucket breaches the threshold.

Severity

Configure Warning and Critical thresholds independently. Both are evaluated on every scheduler tick.

Example configurations

Alert: Notify when web response time exceeds 2 seconds for 5 minutes.

Setting	Value
Metric	Web Response Time
Operator	above
Critical threshold	2 seconds
Duration	5 minutes
Time function	all

Alert: Warn when HTTP failure rate exceeds 5%.

Setting	Value
Metric	HTTP Failure Rate
Operator	above
Warning threshold	5 %
Critical threshold	15 %
Duration	5 minutes
Time function	any

Alert: Database query time exceeds 0.5 seconds.

Setting	Value
Metric	Database Query Time
Operator	above
Warning threshold	0.1 seconds
Critical threshold	0.5 seconds
Duration	10 minutes
Time function	all

​Available metrics

​Web transaction metrics

​Background job metrics

​External service metrics

​Database metrics

​Error metrics

​Rule types

​Baseline metrics

​Targets

​How alert evaluation works

​Operators

​Evaluation windows

​Time functions

​Severity

​Example configurations