I want to understand the best practices of using the one free unattended robot license to manage multiple unrelated processes. Let me explain what I want.
I want to use it as an IT incident adaptive and auto response engine. Suppose we have 20 types of IT incidents that might be triggered separately by email (or other means) at ANY time of a day: A1–A20. It is unlikely they happen at the same time and it can be tolerated to take care of them in sequence if that happens. Most of the response processes require UI interactions. I would like to utilize one robot (constrained by license) to best manage and scale. Can some experts explain to me the pros and cons of the following approaches.
process: 1X giant state-machine/transaction process with internal branches (states invoking separate activities ) for all 20 incidents.
schedule: it will be a loop so it occupies the entire cycle making it hard to run other processes.
This makes maintenance and scaling of this process quite challenging for a multi-person team.
process(es): 20 processes that each take care of 1 type of incidents.
schedule: use orchestrator to schedule them to run 1 after another within 1 hour. So each process will have opportunity to run every hour.
Some of the processes may take more than 3 minutes to run if there is an incident, what happens when it steps on the next process’s time slot?
S3 use high density robot???
processes: 20 processes that each belong to a a different service account.
schedule: Can this work with unattended robot?
S4 use 1 underground process and 20 gui process
processes: 1 trigger process that checks emails and decides what worker process to run (start-job)
20 gui processes, deployed but not scheduled. Will this work???
schedule: deploy all processes but only schedule the daemon process.
S5: use 20 underground processes and 20 gui processes.
I’m very very new, so please be gentle! Thanks!