Large deviations from expected bond lengths reported by gmx check

GROMACS version: 2024
GROMACS modification: Yes

I am not getting any warnings during the simulation run, but when I perform gmx check on the output topology, I get output like this:

Reading frame 3 time 1500.000
Distance between atoms 159 and 161 is 0.145, should be 0.134
Distance between atoms 872 and 884 is 0.163, should be 0.149
Distance between atoms 949 and 951 is 0.148, should be 0.138
Distance between atoms 1240 and 1242 is 0.155, should be 0.143
Distance between atoms 1441 and 1448 is 0.163, should be 0.149
Distance between atoms 1718 and 1730 is 0.160, should be 0.149
Distance between atoms 1859 and 1867 is 0.161, should be 0.149
Distance between atoms 1890 and 1892 is 0.155, should be 0.143
Distance between atoms 2203 and 2207 is 0.126, should be 0.137

Are deviations of this magnitude (>0.1 Angstrom) reasonable, or does this indicate a problem with the simulation?