vRealize Operations - Any input on implementation planning?

[PCguyLearningMA]

I have completed a Proof of Concepts of vRealize Operations Manager.  Now I am building out the project plan for real-live implementation.

Do any folks who have deployed in a medium sized implementation have some insight, gotchas, ah-ha's you would care to share?

Kind Regards

Ben

[mark_j]

Few suggestions:

• Look to the published ref architecture for guidance.
• Deploy RC cluster in a minimum of 2-node collector groups.
• Deploy the load balancer VIP for the analytics nodes for product UI acces (443).
• Deploy load balancer VIPs for remote collector groups for EPO agent connectivity.
• Reserve more DNS/IPs than you need (e.x. couple extra nodes in Analytics cluster and RC groups) - add those DNS names to the SAN cert so you're future-proofed on adding future nodes.
• Factor in a (minimum) 4-week lead-time between installation and go-live, as you need the 3-4 week of historical data to have accurate DTs and forecasting. Consider that 4-week period your burn-in and period for extended configuration.
• Watch your network latencies... stay under 200ms between analytics cluster and RC, and 20ms between EPO agents and RCs.
• Consider the work products such as dashboards, views , reports, alerting, as iterative+progressive (it's not only 0% or 100%) and don't mark them complete until they've been reviewed+approved by the stakeholders.
• When you make the project plan, make the work products of content clearly quantified so you have a measurable goal to denote completion. Some people use # work products, but I find it more realistic to define the use case goals and work towards that. With subjective content like vR Ops content, it can be construed that it is "done  when we say it it". That can become problematic (risky) from a planning perspective because you cannot practically measure a person's warm-n-fuzzies.

[BenJ_UAlb]

Great suggestions.   Thank you!

Kind Regards,

Jamin