Subnets/ranges per discovery schedule

[rgu]

Based on your experience, how many subnets/ranges are too many to include in a single discovery schedule?

I'm sure the answer to that could be, "it depends". But I imagine based on usage of ServiceNow Discovery, you've come up with a ballpark number of subnets/ranges you don't go over in a single discovery schedule.

We have 4000 subnets to scan weekly, with a 7 hour overnight window 5 nights a week. Are there compelling reasons not to put 800 subnets/ranges in a single discovery schedule (assuming we have the necessary number of MID servers to complete the scans in the 7 hour windows)?

(For the purposes of this discussion, assume the subnets are similar in nature and don't need to be split up for other logistical reasons (DMZ vs non-DMZ, geography, etc.).)

(We are aware of the "Run After" option with discovery schedules, but for reasons that would be too long to explain in this post, we would prefer not to use them for this.)

Thanks for your time.

[cyked]

Our schedules are setup to run for 8 hours, every 3 days.   We have three schedules (A B C) between two data centers split up by prod and non-prod.   The schedules are a mix of /21, /23, /24m but mostly /24.   One schedule has 13 networks in it, another with 10.   The others have 5-8.   The larger schedule runs for around 6 hours (at night).   Hope that helps.

So it definately "depends" on the netmask you are defning and how many hosts are actually active on those subnets.  

[doug_schulze]

And the number of midservers you have clustered to achieve the goal..As I say in all the "shows" a single midserver can do a million devices, it would just take a year or two to get through .. you can do whatever you need to accomplish just have the right amount of mids and threads to make it happen