FNMS on-prem migration to Cloud
Does anyone here have any experience of migrating Flexera (FNMS) infra from on-prem to cloud (IaaS)?
We have checked with Flexera Professional Services, and they said they have not had any customers who have tried to use a private cloud infrastructure (azure/google cloud) to install FNMS On-Premises servers.
Looking for some answers on FNMS on-prem performance on IaaS cloud.
Appreciate any reply from the forum here. Thank you.
It might depend on who you ask, but it is common to run FNMS On Premises servers in a private cloud infrastructure provided by the Azure or AWS. I haven't personally heard of any organizations putting their FNMS servers in Google Cloud, but that doesn't mean it hasn't been done.
Flexera One ITAM (which shares many of the same components under the hood as FNMS On Premises) actually runs in AWS. Flexera has been able to get much better performance from the cloud infrastructure than we were able to get using self-hosted and managed infrastructure.
Flexera One is definitely listed in our options, however at this point, the direction is to move the infra to cloud.
Let me give more detail on our issue/concern.
Currently our on-prem DB server is running with NVMe solution with 275k read IOPS/ 325k write IOPS. The card was chosen because we had issue with standard disk few years ago where imports took 1 week to run and it was only 20% completed.
Now, Azure does not offer such high IOPS on their VM, the highest spec (Ultra disk) offers up to 160k IOPS only.
Therefore we are looking for some sharing from the users here, how is the performance of FNMS on-prem in private cloud? Does anyone see any performance issue such as inventory import / reconciliation taking longer time than usual?
I have definitely seen some performance issues in local testing as well as on-prem customer environments lately. Changing database compatibility level back to 110 (2012) has helped, although the used DBMS versions were supported (2017 & 2019). Wouldn't say this is the one setting to defeat all performance problems, but a thing to consider.