Enhancing Error Handling and Checks in MentoV2 SDKs/Swaps

[denviljclarke]

Description

An issue has been identified by zvida, a Celo Terminal engineer, regarding the MentoV2 SDKs and swaps functionality. The problem occurred when the reserve runs out of axlUSDC, causing numerous flows to break. This issue is compounded by the lack of checks for actual values within the registry, leading to failures in swap operations due to insufficient funds in the reserve.

Specific Problems Identified:

1. MentoV2 SDKs/Swaps Failure: The SDKs and swaps do not account for the reserve's actual balance, resulting in broken workflows when there's a shortage of axlUSDC.
2. Confusing Error Messages: The contract does not verify reserve balances until a transfer is initiated, leading to unclear and confusing error messages.
3. Lack of Pre-emptive Checks: Currently, there's no mechanism to check the reserve status in advance, either in the SDK or the broker contract, to prevent such failures.

Proposed Solutions

• Implement Checks in SDK: Integrate a system to verify the actual balance in the reserve within the SDK. This check should be designed to prevent swap operations if the reserve is low on funds.
• Enhance Contract Error Handling: Modify the contract to conduct early checks for sufficient reserve funds before initiating a transfer. This change aims to provide clearer error messages and prevent transactions that are bound to fail.
• Enhance Contract Error Logging: We can decide if and where we would like to add logging using Sentry

This issue is more about exploring error handling and logging as a whole in the SDK.

Action Items

• [ ] Analyze the current SDK code to identify the best integration points for improved error handling.
• [ ] Decide if and where we should add logging

Acceptance Criteria

• [ ] A plan for better error handling if/how it can be improved.

[denviljclarke]

We've decided that this error handling should be done on by the SDK user, however based on this we've also decided it would be good to handle this type of error and provide some feedback to users when this occurs on the front on so I've created another issue to track this.