you can add an event subprocess starting with a timer when the process instance gets overdue. In the event subprocess you can do some activities t remind somebody or take care about the environment. Event Subprocess | docs.camunda.org
But don’t you then try to solve a problem (worker stopped working) with the thing that caused the problem? I mean, in order to send a failure, you need a worker. But if the whole client (with all its workers) is down…
My setup has 1 service that allows users to interact with UI where users can trigger start a process. If that service is unavailable then I will know. About other services, I think there is a possibility to interact with them via actuator info or some endpoint to present the application available (I’m using spring-boot-starter-external-task-client) by HTTP connector.