Feature Request, Anonymous Data Collection To Improve Development
Proposal, Collecting anonymous data to monitor FOG usage and improve FOG based on usage patterns.
By collecting information about the types of systems the FOG server is installed on and the systems that FOG helps to image we can have the data to make decisions about development.
Anonymous info to collect:
Options to collect data on the following, opt out tick box:
Fog Server information
CPU info – Model+cores (model may not be enough to detect VMs with less)
Number of Storage nodes configured
Number of registered hosts
Number of images
Disk Read throughput / IOPS benchmark
Network Interface Speed
Compression Level Set
Unicast Imaging Tasks done in last 12 months
Multicast Imaging Tasks done in last 12 months
Advanced Tasks done in last 12 months
CPU – Model+cores
Disk – Model+size
Fog Server needs to generate a UUID for itself
Fog Server also needs to generate UUID for each client
Once a week the Fog Server packages the data selected and sends it to the reporting server.
What could we use information for?
• What features are people using, which ones should the developers focus on
• What machines do not work with FOG, so the developers can troubleshoot
• What storage and NICs are users giving FOG? Can we improve how we utilize faster NICs and storage.
• Is there spare CPU/RAM capacity on the FOG server or clients for more intelligence or processes?
• Many more which I’m sure will come about due to looking at statistics.
It’s definitely the place to start though.
The code just probably needs updated to work with the current versions.
Should also make sure the receiving server is still operational…
it’s worth mentioning that there has been for a long time an optional (default: off) setting in fog called FOG_ADVANCED_STATISTICS with the description of ```
Enable the collection and display of advanced statistics. This information WILL be sent to a remote server! This information is used by the FOG team to see how FOG is being used. The information that will be sent includes the server’s UUID value, the number of hosts present in FOG, and number of images on your FOG server and well as total image space used.
I, too, like the idea of this, though I have a feeling this is, hopefully relatively soon, going to become something that’s much easier to work within. I totally agree that having more anonymous statistics could help us, but it also means more users using their bandwidth.
As far as if it’s easy. This isn’t really something that is easy, or totally transparent. The amount and type of data to be requested in such a manner requires a lot of finesse and tweaking to get only the information we care about without a potential pitfall of a security issue.
I like the idea, but it might be easier to pull anonymous data from the FOG database first. Then work on implementing the other stuff, like FOG server # of cores, RAM, etc.
Unless that stuff is super easy to pull… I haven’t looked into it.