Usually what I do in these types of automations is include a second 'failsafe' script activity after the one used to initialize the next automation. I have found that depending on the activity in the account, a start automation API request in a constant loop can be 'lost in the nether' and although infrequent, can completely screw up your whole process.
So basically you would have:
Your Actual Actions | Script1 | Wait 2 minute | Script2
- Script1 is your current 'start next automation' script
- Wait period is to ensure the request is complete or any 'starting weirdness' is gone. I put 2 mins, as that was acceptable to my Use Case, but can be 1 minute or even to be extra safe 3-5 mins.
- Script2 is a script to check current status and if it is not
'Running' then to push start request again.
This failsafe has reduced the amount of times that the automation loop has grinded to a halt to once or twice every year.
If you really want to keep on top of it, you can also set up a third scheduled automation that retrieves the status of the 2 other automations and if neither one is 'Running' then to send an alert email for you to check it out. You could also have this automatically start it again, but I tend to like the manual check to verify it and will be helpful if there is an error as well.