Am looking for difference between OPM(OCUM7.1P1) and Grafana. we have plan to suggest Grafana in our environment and there was questions to justify when OPM is available, why looking for grafana. with my experience, i can say that Grafana gives more granula data which gives more detailed at every layer. But looking for some additional details so as to convince and get in our environment.
we have 7mode and Cdot filers with versions starting from 8.2.3 to 9.1.
you delivered the strongest argument yourself. OCUM is cDOT only. If you want to collect performance data from 7mode, you'd need to setup Unified Manager for 7mode. Two seperate tools to collect and analyze performance data. No single source of information, different level of information, different UIs, different workflows.
Grafana will get all data in one place. Single tool, single source of truth.
As Niels rightly pointed out, UM 7.x/9.x is for cDOT environments only so if you have both 7mode and cDOT you will need UM 5.x for 7mode and 9.x for cDOT instance.
Having said that NetApp has invested a lot in Unified Manager 7.x/9.x analytics for providing cooked performance and capacity data in the form of reports, graphs and charts which are based on algorithemic calculations making it easier for users to understand ONTAP performance and capacity. To go along, the events come with auto-remedial script functionlity which be leveraged for common to complex usecases.
Given that Harvest is community supported (although backed by a vast respondents) you have to pick the right software wisely. Feel free to reach out to me if you need further details on UM
In my opinion, it is not choice A or B. Why not having them both.... When looking a harvest/grafana there is not a big investment....you only need an appropriate VM.... Even more the harvest/grafana can also be extended by 7-mode systems and other interesting environments like VMware..... Might be interesting to map similar statistics from VMware with Netapp.... 🙂
You are right but there are other aspects of UM usage which is not covered in this discussion, UM provides various analytics to easy the day-to-day life of most engineers, with:
1. Failover Planning: You can use data and graphs to compare between systems and plan for failover
2. Workload rebalancing: Use performance capacity and utilization metrics to plan and implement workload rebalancing
3. Performance baselining
4. Inactive data reporting
5. Performance capacity planning
6. Bully victim anaysis
7. Latency charts with detailed delay centers to ensure administrators spend lesser time in troubleshooting storage issues
8. Detailed event analysis with recommended redmedial action plan
9. Auto remediation with alert scripts
10. QoS and Adpative QoS monitoring
11. Various processed metrics using various algorithms to provide users with detailed analysis of ONTAP - Avilable IOPS, IOPS/TB, Performance Capacity used etc.
12. Real-time and polled data for mission critical objects and ONTAP objects in general
13. Protection monitoring with Topology support
These are some salient features apart for event, inventory and capacity analysis that UM provides. Its one thing to support multiple metrics from ONTAP for advanced users (includes some tweaking) and another to provide processed metrics in a easy to understand format that helps almost everybody. Just wanted to highlight this for anybody before making a decision or using the two together.