Large uncertainties despite a good fit

[Natilus]

Dear experts,
I have a Roofit code that fits my data to a combination of a triple-Gaussian (signal) and a straight line (background). Despite having a good fit (good Chi2/ndf, covariance matrix quality, and fit status) and correct yields, the uncertainties on the yields is very large.

I am using Root 6.30/06 on RHEL 9.0.
I am uploading my code with one of the fits I performed in case it helps.
The main fitting code is called “FitMuMuResults.h”. I also included the header file where I define my class.
Some statistics:
Chi-Squared over Number of Degrees of Freedom = 1.108059
Covariance Matrix (Fit) Quality = 3.000000
Fit Status = 0.000000

Thank you in advance for the help

FitMuMuResults.h (19.3 KB)
FitResults.cc (3.8 KB)
RooFitter.h (6.3 KB)
FitEMuResults.h (16.7 KB)

[couet]

I think @jonas can help you, but maybe you should first provide the right include file because when I execute your macro I get:

root [0] 
Processing FitResults.cc...
In file included from input_line_8:1:
FitResults.cc:3:10: fatal error: 'FitEMuResults.h' file not found
#include "FitEMuResults.h"
         ^~~~~~~~~~~~~~~~~
root [1] 

[Natilus]

Dear @couet ,
Thank you for the quick response,
I uploaded the other file too.
FitEMuResults.h is also a fitting code similar to FitMuMuResults.h but simpler.
If I can find a way to fix the problem with one of these codes, I can do the same with the other.

Thanks again for the help