Cannot create trigger from anomaly detector - requestTimeOut

Plugins Alerting
1

Hello
I’m experiencing request timeout when I try to create a trigger for my monitor that was defined using anomaly detector in Kibana.

I set Trigger type as ‘Anomaly detector grade and confidence’, and for a few seconds it tries to fetch some data from detector in ‘Define type of trigger’.

And it finally says this:
“No features have been added to this anomaly detector. A feature is a metric that is used for anomaly detection. A detector can discover anomalies across one or more features.”
when my detector does have features set up, and is already running and detecting anomalies.

I looked into the logs in pod for kibana(i used helm to install elasticsearch), and there is this log:

Alerting - AnomalyDetectorService - getDetectorResults { Error: Request Timeout after 30000ms
at /usr/share/kibana/node_modules/elasticsearch/src/lib/transport.js:397:9
at Timeout. (/usr/share/kibana/node_modules/elasticsearch/src/lib/transport.js:429:7)
at ontimeout (timers.js:436:11)
at tryOnTimeout (timers.js:300:5)
at listOnTimeout (timers.js:263:5)
at Timer.processTimers (timers.js:223:10)
status: undefined,
displayName: ‘RequestTimeout’,
message: ‘Request Timeout after 30000ms’,
body: undefined }

Has anyone experienced same problems as mine? I’d appreciate any suggestions. Thanks

2

Alerting create AD trigger page will call preview detector API to get sample anomaly results and detector feature configuration. The preview API may not work for some cases

  1. If preview detector API timeout or throw error, AD trigger page can’t get any data, then it will show this warning “No features have been added to this anomaly detector. A feature is a metric that is used for anomaly detection. A detector can discover anomalies across one or more features.”. The message is not so accurate for such case as detector have features but just failed to call preview API.

  2. When add AD trigger, alerting frontend will call preview API with last 5 days data. And preview API needs 400+ data points to generate sample anomaly results. If the anomaly detector interval is too big or there is no enough data in last 5 days, then preview API won’t return anything, then we will see this error message.

We will tune this error message. You can ignore this error message now as this is just from preview detector API, actually the detector job can still run correctly at backend, just some wrong error message on UI. You can still create monitor and later you can tune the monitor with custom query.

Suggest to try this preview API on Kibana DevTools to check if preview API can return any result, Anomaly Detection API - Open Distro Documentation

POST _opendistro/_anomaly_detection/detectors/<detector_id>/_preview
{
  "period_start": 1636147713000,
  "period_end": 1636579713000
}