Is it normal to get 3000+ vulnerabilities while scanning server using Nexpose? Or I can reduce the vulnerability to decent count by some tweaks!
Is it normal?
That depends entirely on the server in question. If it's an ancient machine running ancient software, then I would not be surprized to find thousands of thousands of vulnerabilities. If it was a completely new server set up according to best practices with barely any content on it - then yes, that would definitely be abnormal.
How can I reduce the number of vulnerabilities?
The painfully obvious solution would be to fix the vulnerabilities. You have bought a vulnerability scanner to show you vulnerabilities for a reason, after all.
Can this be tweaked somehow?
Yes. Vulnerability scanners generally have a sense of how "certain" they are that a finding is actually a vulnerability, as opposed to a "maybe could possibly be a vulnerability, perhaps". As certainty decreases, you will more and more likely run into findings that are False Positives - meaning the scanner detected something that is not actually a vulnerability. What you care about are so-called True Positives, meaning the scanner identified something that is actually a vulnerability.
You may now be tempted to turn down the sensitivity of the scanner to reduce the amount of False Positives and leave you with a high density of True Positives. However, this will inevitably increase the number of False Negatives, meaning the scanner did not identify something as a vulnerability, even though it is really there.
How to proceed now?
Whatever path you choose, it will either result in a lot of work, or in a high risk for you.
The best possible strategy is now to create a matrix of vulnerabilities, with one axis being the certainty of a vulnerability, ranging from "100% confirmed" to "possibly maybe", and the other axis being the severity of the vulnerability, ranging from "Your server is on fire" to "Could possibly be inconvenient". Here you see a little diagram of what that matrix could look like:
High Severity, High Certainty - These findings need your immediate attention, as they have the most potential to be exploited by attackers.
An example would be an SQL Injection in your web application, allowing attackers to read and modify all data, and perhaps even execute arbitrary code on your server.
High Severity, Low Certainty - These findings should be examined afterwards and determined if the finding actually constitutes a vulnerability or not.
An example could be outdated software on your server with publicly available exploits, determined by version number. This could be misleading, since maintainers may offer security backports, which fix the vulnerability, but do not change the version number. If you are however affected by the vulnerability, attackers may be able to execute arbitrary code on your server, depending on the vulnerability in question.
Low Severity, High Certainty - While these findings are very clear they exist, they don't pose a substantial risk to your server, and thus should be patched "later down the line". Just make sure that it actually happens and that "in a year" will not be perpetually a year away.
An example of this could be configuration flaws for TLS (e.g. support for TLS 1.0), missing security headers, disclosure of version information, etc.. While the scanner can confirm that with high certainty, the impact of these is generally low enough to not require immediate action.
Low Severity, Low Certainty - These findings can usually be ignored. It's not certain if they are actual vulnerabilities to begin with, and even if they were, their impact would be rather low anyways.
I'm struggling to think of a good example here, since they don't really matter that much anyways.
You could, of course, also go down the path where you reduce the sensitivity until you have only a handful of findings, then have your IT tell you that they won't fix them, and finally everybody successfully wasted a lot of time and money and did absolutely nothing to improve security whatsoever. As you may be able to tell, this is not advisable to do - but it is nonetheless the route companies choose to go down. These companies then find themselves in the news for massive data breaches.
What about "Informational" findings?
A lot of vulnerability scanners also include "Informational" findings, which are things like a port scan, a list of ciphersuites supported by the server, etc.. These are not vulnerabilities and don't have a direct impact on you. They're just information you may or may not be interested in.
Compare this to a log file, telling you when a server was started and stopped - it's good to have this information in case you need it, but it doesn't require any actions.
It depends, but sometimes to note:
Not all of the found vulnerabilities may necessarily be real. Like AVs and IDSes, vulnerability scanners can have false positives, which is when it detects something when nothing is there. If there is a case of having too many false positives, then you need to tune the system.
We all know the classic story about a doctor testing someone positive for cancer, when the person had no cancer in the first place.
Saying a vulnerability scanner found something is one thing and proving it is another. This is where we will perform vulnerability validation to determine if it exist or not and if it does what impact will it have.
How can I know if a found vulnerability is real?
You can test it by trying to exploit the vulnerability and see if it works. This is one of the reasons people do stuff like pen-testing.
Side note: Be careful when it comes to a system or application that is mission critical or have a very important function. You don't want to break anything.