Overview

Metric monitors offer enhanced options for handling missing data, allowing you to differentiate between missing data as a failure mode and a healthy state.

These options align with what’s available in other monitors types like Logs, Events, CI, Database, Error Tracking, and more.

Benefits of using On Missing Data options

When measuring the number of bad events, like errors, Monitors should reflect an “OK” when no data is detected. With the legacy No Data configurations, monitors would report No Data. The On Missing Data configuration options allows monitors to reflect health states more accurately, improving clarity.

Monitors managed through the UI

If you manage your monitors from the UI, the configuration automatically updates the next time you edit them. To update the On Missing Data configuration sooner, see the following sections on adjusting through the API.

Monitors managed through the API or Terraform

If you are managing your monitors with API or Terraform, replace notify_no_data and no_data_timeframe with on_missing_data. The no_data_timeframe parameter is not required since on_missing_data uses the same time frame as the time window.

API parameters

The previous No Data parameter, notify_no_data, remains available on existing monitors and are not automatically upgraded to the new on_missing_data features.

For all the available fields, see the API Documentation.

Here’s an example of before and after of a JSON monitor with those fields:

Before

{ 
  "name": "CPU usage is high for host $host.value",  
    "type": "query alert",  
    "query": "avg(last_5m):100 - avg:system.cpu.idle{$host} > 90",  
    "message": "A high CPU usage has been detected for host $host.value, which can impact the system performance.",  
    "tags": [],  
    "options": {  
        "thresholds": { "critical": 90 },  
        "notify_audit": false,  
        "include_tags": false,  
        "notify_no_data": true,  
        "no_data_timeframe": 10  
    }  
}

After

{
  "name": "CPU usage is high for host $host.value",  
    "type": "query alert",  
    "query": "avg(last_5m):100 - avg:system.cpu.idle{$host} > 90",  
    "message": "A high CPU usage has been detected for host $host.value, which can impact the system performance.",  
    "tags": [],  
    "options": {  
       "thresholds": { "critical": 90 },  
       "notify_audit": false,  
       "include_tags": false,  
       "on_missing_data": "show_and_notify_no_data"  
    }  
}  

SLO monitors

SLOs treat uptime and downtime according to this mapping:

Further reading

추가 유용한 문서, 링크 및 기사:

Monitors API DocumentationAPI