Problem calling eth_chainId on mainnet

[danjm]

On 4.17.0, there is a bug that causes MetaMask to take an extremely long time to connect to mainnet.

One public report of the bug can be found here: https://twitter.com/poopie_cat/status/1058786439993806848?s=21

The bug was introduced by #5552, which was reverted #5662. This revert was a hotfix, published with 4.17.1

The cause of the bug is not the code change in #5552 directly, but rather an interaction between that code change and infura.

For some reason, the GET https://api.infura.io/v1/jsonrpc/mainnet/eth_chainId?params=%5B%5D request initiated by the ethQuery.sendAsync({ method: 'eth_chainId' }, (err, chainIdHex) => { call found at https://github.com/MetaMask/metamask-extension/pull/5552/files#diff-684dc3d1f34ef1201f93c566be5530abR111, takes very long time to resolve on mainnet, but not on other networks.

On other networks I am seeing it resolve in ~100ms, with a 400 error. On mainnet it takes 2 minutes or more. And eventually resolves with a 503 error:

If we log the error passed to the callback from the above referenced line of code, we see this error:

Error: InfuraProvider - cannot complete request. All retries exhausted.
Original Error:
Error: Gateway timeout. The request took too long to process. This can happen when querying logs over too wide a block range.
    at createAsyncMiddleware (index.js:40)

It is unclear whether this bug can be resolved within MetaMask's provider or JSON rpc related dependencies, or if it needs to be addressed by Infura.

Please comment if you have further insights or questions.

[kumavis]

We can hardcode the response to eth_chainId for each network

[tejaslende]

Please check out this web3auth https://web3auth.io/docs/connect-blockchain/bnb