We did have that some time ago but we did remove the 'All' option. Its a very dangerous setting in large environments.
Also the benefit is not really clear as those services are completely different in terms of response times and error rates and an 'all' threshold in most cases does not make much sense.
What would you like to do?
What I wanted to do was to just have a notification sent anytime an error (HTTP / Server side error) was detected for any service. As you said above it probably would not make sense to have such an alert in large environments. In our case, it's nice to have an alert that an error was detected on a service / multiple services.
Similarly it would have been nice to have a general error metric that stands for any kind of error. Users can choose this if they want to track all errors (instead of having four categories such as Client side, Server side, HTTP 4xx, HTTP 5xx). So, basically it could easily cover the use case "For any error detected on any service let me know". Based on your answer, I think my only option is to set it up manually for each service.
The thing is client got alert after 9 minutes when one process got connectivity rate 0%. Problem here that ops team were notified than something isn't working by managers before Dynatrace did it. I know we have sliding 5 and 15 minutes interval, but maybe there is an option to get notification much faster.
@Vlad S. Why did'nt you got a much faster service related alert? We are alerting not necessarily on 5 min delay, if your service has high traffic and you get an increased error rate you will receive the alert within a minute instead of 5 minutes.