Problem with metrics built-in

[gustavodutra]

Hi, everyone,

Has anyone here experienced issues with AWS Lambda built-in metrics?

Recently, I noticed that the Lambdas configured in my AWS environment have been failing frequently, and Dynatrace received the error metrics but did not generate any alerts.

Dynatrace has the capability to connect with the AWS environment and bring insights into the tool. It works reasonably well, but it’s usable.

Has anyone else been through this?

[dawid_kaszubski]

Hi @gustavodutra,

Dynatrace, by default, triggers an alert when the failed invocations rate is higher than 5%. 

You can check the alert configuration in "Anomaly detection for classic AWS services" tab in Settings classic. 

[gustavodutra]

Hi, @dawid_kaszubski.

How are you? 

I don't find this option, could you send me an example screenshot or the referenced URL?

[dawid_kaszubski]

You can find the "Anomaly detection for classic AWS services" page, by clicking "AWS" in the "Anomaly detection" section. 
URL: 

<the tenant url>/ui/apps/dynatrace.classic.settings/ui/settings/builtin:anomaly-detection.infrastructure-aws

[gustavodutra]

I found the option below in automatic mode and changed it to custom settings, setting a threshold of 5%; however, the lambdas aren't triggering alarms.

I have two lambdas that have 100% failure rates and aren't generating any alerts.

[gustavodutra]

Guys, I managed to get around this by increasing the sliding window and dealership sample.

I set both to 60, and the alerts appeared, and they seem to be helping.