The bootstrap capacitors (C1, C5) in that circuit are very large and there is no current limiting
resistor in the bootstrap path, which is distinctly odd.
The resistors R3,4,7,8 are entirely redundant, the gate driver chip handles everything to do with
shutdown. I can see that these have motivated increasing the bootstrap caps to much larger values
than normal. The website that circuit came from has some unconvincing explanation of why those
resistors are there (*), but I’ve never seen or used that before, it seems to be some protection scheme
for high voltage use.
Standard bootstrap caps are about 10 to 20 times the total gate capacitance. You can add 10 ohms
or so in series with the bootstrap diodes to reduce current pulses through the diodes to a sensible
Make sure the decoupling for the IR2110 itself is 10x or more than the bootstrap caps.
All 12V decoupling ceramic. Careful attention to ground plane / layout is needed (the output traces
and bootstrap components that swing at high dV/dt should not be over a ground plane, note.)
(*) I think the originator was using a high voltage bridge without powering up the 12V before the main
high voltage supply - that is always likely to lead to failure - in fact with high power circuitry a soft-switch
is normally used to prevent massive inrush currents to high voltage decoupling electrolytics.