The iterations should be the maximum supported by the computer, giving maximum tolerable execution time.
The NIST already recommends uses 10,000 iterations:
For PBKDF2, the cost factor is an iteration count: the more times the PBKDF2 function is iterated, the longer it takes to compute the password hash. Therefore, the iteration count SHOULD be as large as verification server performance will allow, typically at least 10,000 iterations.
Because SHA-1 is way faster than SHA-2. The SHA-3 (Keccak) is slow in software, but faster in hardware than SHA-2. So, stick with SHA-2 can be more secure against FPGA.
I think replace the PBKDF2 to Argon2id, the winner of PHC, can be better, but will need more changes in code.
The PBKDF2 is faster than Argon2id or BCrypt. It will be even faster with low quantities of iterations. By default it uses:
https://github.com/AugustoResende/RaiLightWallet/blob/master/src/js/rai-wallet/Wallet.js#L167
The LastPass uses 100,000 iterations using SHA-256. The Covecube uses 200,000 iterations using SHA-512.
The iterations should be the maximum supported by the computer, giving maximum tolerable execution time.
The NIST already recommends uses 10,000 iterations:
https://pages.nist.gov/800-63-3/sp800-63b.html#sec5
Also, it seems better change the SHA-1 to SHA-2 (SHA-512):
https://github.com/AugustoResende/RaiLightWallet/blob/master/src/js/rai-wallet/Wallet.js#L1143
Because SHA-1 is way faster than SHA-2. The SHA-3 (Keccak) is slow in software, but faster in hardware than SHA-2. So, stick with SHA-2 can be more secure against FPGA.
I think replace the PBKDF2 to Argon2id, the winner of PHC, can be better, but will need more changes in code.