Throughput measurement for services that call the Database

[yuval1983]

Hello everyone!

We currently have hundreds of services in the system. If there is a significant increase in calls to the database, we need to receive an alert. What is the best way to accomplish this?

is an example. After service, there were many calls  we would like to receive an alert (anomaly, of course)

[marina_pollehn]

If you want to keep it relatively standardized and use the load baseline generated by Dynatrace, I would suggest activating the detect load spikes setting for the database service.

You can find it when you select the database service and click on the three dots in the top right. Go to Settings and activate the load detection in the anomaly detection setting (scroll down). It is turned off by default.

Note, this is monitoring the requests to, not the requests from so you will be alerted with a generated Dynatrace problem if the number of database statements increases drastically (you can adjust the sensitivity with a relative threshold as shown in the screenshot).

If you rather want to use a fixed threshold or monitor the number of requests from, I suggest working with a custom alert (metric event).

Use the number of calls to databases as a metric key. 

[yuval1983]

HI,

Why can't I work if number of calls to databases?

I actually want an anomaly and then receive an alert on the subject with the help of a metric or Problem alerting profiles

builtin:service.dbChildCallCount:splitBy("dt.entity.service"):sum

[yuval1983]

HI

builtin:service.dbChildCallCount:splitBy("dt.entity.service"):value:default(0):rate(1m):setUnit(PerMinute):sort(value(sum,descending)):limit(20)

How do I change the result to percentages?

and

((builtin:service.dbChildCallCount:splitBy("dt.entity.service"))/(builtin:service.requestCount.server:splitBy("dt.entity.service")))

Is there anything more true?

Is it possible to do something more precise?