Failure rate/response time problems on Services as a whole rather than on Requests

[yair_mashmor]

Hi all,

Wanted to know if anyone else wondered about this or managed to do this -

We have services that have under them many dynamic requests.

When one of these requests is experiencing a failure rate increase or response time degradation, the problem in dynatrace is opened on the service itself.

We would like that to change and only open a problem when the failure rate or response time degradation has change on the service as a whole (after all these metrics are measured on the service itself also and not only on individual requests).

We do not want to exclude these requests from being monitored or to set static baselines.

can this be done? has someone managed to do this? is this is in planning?

We are currently being bombarded with problems being opened in the system and cannot really focus on the important ones.

Thank you,

Yair

[Babar_Qayyum]

Hello @Yair R.

You are right but for the time being, I think it is not possible because request/s belong to the service, and the configurations available only on the service level.

I think you should reconfigure the services detection sensitivity, set alert thresholds, or disable alerting for certain services (not important).

Regards,

Babar

[yair_mashmor]

Thank you Babar.

Would love if maybe Staff could way in and verify if this is possible or not (or others that might have an answer from support).

Thank you

[qxjl]

we have not try this but I guess this could be achieved with Merged service , you can merged all of your services as one , then it would use the same set of metrics for all your requests .

[Yosi_Neuman · Accepted solution]

Hi @Yair R.

Long time no see 🙂

If you mean that on one service with lots of different requests under it you wants to get alerts if the entire requests start to fail and not to get alert if only one or tow request start to fail, you can do the follows:

1. Shut down the anomaly detection for errors on the service level.

2. Create error metric with the MDA on the service without splitting by requests.

3. Set create custom alert on the carted error metric.

HTH

Yos