Check the logs, is it waking, sleeping, waking, sleeping, waking, sleeping? Power management bugs are super annoying because they are tedious to dig into.
In my very obscure case, I got lucky in that suspend was working, then stopped working after a kernel update. Since it was a regression, kernel developers were more interested in tracking down the problem and I was able to find a work around: write PWRB to /proc/acpi/wakeup
The gory details are here, which I expect is so obscure it's not your problem, but shows as tedious as it is, filing bugs with a decently good bug report and willingness to do the work devs need you to do can be worth it.
https://bugzilla.kernel.org/show_bug.cgi?id=185521
In my very obscure case, I got lucky in that suspend was working, then stopped working after a kernel update. Since it was a regression, kernel developers were more interested in tracking down the problem and I was able to find a work around: write PWRB to /proc/acpi/wakeup
The gory details are here, which I expect is so obscure it's not your problem, but shows as tedious as it is, filing bugs with a decently good bug report and willingness to do the work devs need you to do can be worth it. https://bugzilla.kernel.org/show_bug.cgi?id=185521