Performance: Logs - Metrics for createKey and Import

[cmgustavo]

Performance Analysis – createKey & serverAssistedImport

This PR adds instrumentation and analyzes app-side performance for wallet creation and import flows.

Results

createKey (11 wallets)

• Total: ~36.6s
• createMultipleWallets: ~29.6s (~81%)
• BWC.createKey: ~7.0s (~19%)

serverAssistedImport (36 wallets, 14 tokens)

• Total: ~77.1s
• serverAssistedImport: ~64.0s (~83%)
• createWalletsForAccounts: ~12.9s (~17%)

---

Key Findings

• Bottlenecks are not UI-related (Redux/object building negligible)
• Main costs:
  • Wallet creation (createWallet)
  • Address generation (createWalletAddress)
  • Server-assisted import scan
• Execution is mostly sequential (~2.7s per wallet)
• Import has wasted work (COPAYER_REGISTERED retries)
• Post-import address hydration can take 14–17s per wallet

---

Optimization Targets

1. Stream import results to unblock UI
2. Avoid duplicate wallet creation attempts
3. Parallelize wallet + address creation where safe
4. Defer/lazy-load address generation
5. Investigate BWC.createKey (~7s)

[cmgustavo]

createKey took 36.65s end-to-end for 11 wallets on livenet.

Biggest bottlenecks:

• createMultipleWallets.TOTAL: 29.65s (80.9%)
• createKey.BWC.createKey: 6.97s (19.0%)

Per-wallet average:

• ~2.70s per base wallet flow

Main repeated costs:

• createWallet.TOTAL: generally 1.45s–1.98s
• createWalletAddress.TOTAL: generally 0.86s–1.61s

Worst address-generation cases:

• XRP: 1.61s
• ETH: 1.33s
• MATIC/BASE/OP/ARB: about 1.23s–1.25s

Negligible:

• buildWalletObjectsAndSubscriptions: 66ms
• buildKeyObj: 1.4ms
• success dispatch: 28ms

Conclusion: the real app-side optimization targets are root key generation, sequential wallet creation, and address generation. Redux/object-building is not the problem.

[cmgustavo]

serverAssistedImport import run took 77.11s end-to-end.

Biggest bottlenecks:

• serverAssistedImport.API.callback: 64.02s (83.0%)
• import.createWalletsForAccounts: 12.86s (16.7%)

Imported result:

• 36 wallets
• 14 token wallets

Negligible:

• dedupe: 0.1ms
• token linking: 13.4ms
• key matching: 0.8ms
• object build: 196.1ms

Important note:

• follow-up account wallet creation is hitting repeated bwc.ErrorCOPAYER_REGISTERED failures for ETH/MATIC/ARB/BASE/OP/SOL
• after import completes, there is also very expensive getLatestMainAddress hydration on many wallets/tokens, which likely contributes to UI still feeling busy

Conclusion: the real import bottlenecks are the server-assisted discovery scan first, then post-import account wallet creation. Token linking and object-building are not the problem.

[cmgustavo]

Fixed: response when click on Create Key (during onboarding)

New changes, commit: 3ae78eb

Controlled concurrency in createMultipleWallets reduced createKey.TOTAL from 36.65s to 20.70s for 11 wallets.

That is:

• ~15.95s faster
• ~43.5% improvement
• ~1.77x speedup

This strongly validates that the previous sequential wallet creation flow was a major bottleneck.

[cmgustavo]

Import improved from ~77.5s to ~66.1s for 36 wallets after removing duplicate wallet creation work. 05ee108

Next optimization step is to defer and chunk wallet address/funds scanning so import can complete earlier and wallet hydration can continue in the background.