walletconnect / blockchain-api

WalletConnect's Blockchain API

WalletConnect Specs

This repo runs specs.walletconnect.com.

Source can be viewed in docs directory.

Running Locally

1. yarn install
2. yarn start
3. Open http://localhost:3000/

Adding a New Page

1. Create a new file in docs/
2. Configure sidebars.js to create the sidebar link.

Editing a Page

1. Open the file in docs/

Older versions are available in versioned_docs/.

Customize Settings & Theme

1. Open docusaurus.config.js

More

For more options, check the Docusaurus docs.

Archived v1.0 docs can be viewed in the v1.0 branch.

blockchain-api's Issues

Install RPC Quotas

Needs design review first https://docs.google.com/document/d/1qfMU7adjKhaE1ynLz4tY_dIhnf5P_gaywa9ZyjXu4do/edit?usp=sharing

Analytics File Thresholds

Currently we roll up files based on:

• file size
• file age

We roll it up into // partitions.

Currently we may not honor these correctly as file size/file age doesn't necessarily match the day.

Hence, we need to introduce another threshold which is when a new day begins.

Add project id validation to the rpc-proxy service

Next steps

• #87
• #81
• Expand to us-east-1 and ap-southeast-1
• #90
• #29 Failover mechanism
• #67 [blocked] Quotas
• [ ]

Track Project ID Cache Hit Ration

We recently had a regression where the project id cache hit ratio went to 0 (Slack thread).

In order to faster see this in the future and to be able to alarm on it we should track the cache hit ration.

AC:

• Metrics added in relay
• Widget added to dashboard
• Alarm defined

Make Sure Rate Limits from RPC providers are tracked as 500 errors to the user

If we get a throttled error / rate limit from an RPC provider we should show that as 500 to a user.
Helps us for metrics and also is the right thing to do.

ECS memory usage creeping up

After each redeploy memory usage starts creeping up, probably there's a leak. Needs investigating

RPC Request stuck

Report back on https://discord.com/channels/492410046307631105/1049240184006381619/1049240184006381619 when fixed

Support multiple providers per chainid

We will soon have multiple enterprise partners per chain id (Infura/Coinbase cloud). Naively let's just round-robin between those providers.

feat: move from warp to axum

Migrate the project from Warp framework to Axum for better scalling and easier maintenance.

Fix tests failing randomly

Integ test seem to fail randomly, probaly due to some race condition.
Works locally, break on github actions.

Needs investigation.

Better error messaging for downstream RPC failures

Action items:

• Attribution: Aggregate 5XX from RPC into HTTP502 response from our RPC
• User confusion: give a good error message to end-users that our chosen RPC provider is down as the reason of the HTTP502
• Failover: give another RPC provider a chance to satisfy the request
• Smart Routing: route traffic to alternative RPC if one RPC is detected to be down

Track NLB target resets

We track metrics from the application level hence might miss the Load Balancer picture.
We use an NLB so it's not a straight up metric.
To improve that and get insight we should track NLB target resets.

Scale down Redis Cluster

Expand dashboard

• Healthy hosts
• What else?

Deployment is broken

Install Usage Quotas for Users

Users should be notified with a path to upgrade when they get close to their quota.

Roll out a project id based allowslist

• Extract the "registry API" client to a separate crate and repository
• #33
• #34
• #40

Add monitoring for RPC's Redis instance

WalletConnect does not currently support the Chinese region?

Describe the bug
If you use a VPN, you can use it. Without a VPN, you cannot use it.

SDK Version

• Version [2.0.0-rc.1]

Screenshots

add `origin` to tracked rpc_requests data

We want to know where are rpc_request originating from so we want to add the origin (eg. opensea.io) to the data we send to the data lake. This will enable us to have more oversight of RPC usage from the different projects or bad actors.

AC:

• a new field is added to the rpc_request data containing origin when available

Gate Prod promotions with integration tests

https://github.com/WalletConnect/rpc-proxy/blob/master/.github/workflows/cd.yaml#L97-L107

We should replace this with the integration tests we have now. they should use the actual staging endpoint of the proxy.

Export Analytics to Athena

In rs-relay we have logic to "dump" certain data (socket/messages) into parquet files. Based on certain thresholds e.g. time/file size we upload these files to S3. From there on our Athena pipeline kicks (in data-lake repo).

Subject of this task is only the file creation and upload part.

feat: add memory and cpu metrics to grafana

Set up a shared redis for the proxy nodes in Terraform.

chore: add binance smart chain support.

Hi! I'm having a problem using WallectConnect with wagmi. Here's my code:

`
const chains = [bsc, bscTestnet];

const { provider } = configureChains(chains, [
walletConnectProvider({
projectId
}),
]);

const wagmiClient = createClient({
autoConnect: true,
connectors: modalConnectors({ appName: "web3Modal", chains }),
provider,
});

const ethereumClient = new EthereumClient(wagmiClient, chains);
const web3modal = new Web3Modal(
{ projectId, network: 'binance' },
ethereumClient,
);

const account = getAccount()

const getAllowance = async (address) => {
const data = await readContract({
address: marketContractAddress,
abi: abi,
functionName: 'getAllowanceToBuy',
args: [address]
}).catch(err => console.log(err))
console.log(data)
return data
}`

It connects and gets account data just fine, but when I call getAllowance I get an error sayng We don't support the chainId you provided: eip155:97
Which is kinda weird, since both WC and Wagmi are supposed to support both BSC (56) and BSC testnet (97).
If anybody knows what to do with it, please, help!

fix: service responds 500 when it should be 401

curl -v -X POST -H "Content-Type:application/json" "https://rpc.walletconnect.com/v1?chainId=eip155:56&projectId=8e6b5ffdcbc9794bf9f4a1952578365b" -d '{"method":"eth_getBalance", "jsonrpc":"2.0", "id":"42", "params":["0xce93f04e99824ba2c56ce20b27ea391abe961bf0", "latest"]}'

Works

When I change the project Id

curl -v -X POST -H "Content-Type:application/json" "https://rpc.walletconnect.com/v1?chainId=eip155:56&projectId=7e6b5ffdcbc9794bf9f4a1952578365b" -d '{"method":"eth_getBalance", "jsonrpc":"2.0", "id":"42", "params":["0xce93f04e99824ba2c56ce20b27ea391abe961bf0", "latest"]}'

I receive HTTP500

Expectation:
HTTP401 or HTTP404

Skip project id validation when /internal API calls

ENS Caching

• Draft ENS Caching docs
• Consult draft version
• Brainstorm the solution on team meeting
• Implement simple Caching
• Implement additional logic

Tasks

Beta Give feedback

1. Initial ENS resolution endpoint rpc-proxy#181
   accepted +1
   
2. Better ENS resolution error handling rpc-proxy#182
   accepted +1
   
3. ENS metrics rpc-proxy#191
   accepted +1
   
4. ENS caching

Update Actions to follow new project specs

Tag images with with versions

Each time we merge app logic (as opposed to terraform) we should build an image with a specific tag such that we can pin this tag in deployments and roll back etc.

We do this in other repos e.g. rs-relay

Expose data in Preset

Acceptance criteria:

• Dashboard depicting the RPC usage exposed in preset (@chadyj can you advise?)

Prerequisites:

• AWS Glue crawler setup in data-lake repository to create Athena tables / maintain Athena partitions.

fix: broken opsgenie notifications

Grafana seems to have an error with opsgenie notifications.

Update privacy policy with links to RPC providers

WalletConnect is careful about what data we collect, but some of the RPC providers we use may be less privacy-focused and log IP address and other PII data. (such as Infura logging IP addresses.

To be transparent we need to update out privacy policy listing our RPC providers, and link to their privacy policies. We should always consider the customer and end-user privacy.

Please list the chains/providers and I will update the policy in https://github.com/WalletConnect/explorer-temp

Implement Proper Integ Tests

Tests should make a random call using all chains we support ideally

Implementation suggestion:
I think the tests in https://github.com/WalletConnect/registry-api-cf-workers/tree/master/integration are a good example

Add WS Support

bug: "The method eth_sendTransaction does not exist is not available"

When I try to call a contract method that requires a user signature, an error is returned to me:
On bsc(id:56) - code: -32000, message: "unknown account", No transaction comes to the wallet.
On mainnet(id:1) - code: -32601, message: "The method eth_sendTransaction does not exist is not available", The transaction does not come to the wallet.

{"jsonrpc":"2.0","id":43,"error":{"code":-32000,"message":"unknown account"}}

I expect that the transaction will come for confirmation in the connected wallet, after confirmation and payment of the fee, the transaction will be successful

async function approveRouter(address) {
const tokenAddress = "0xbb4cdb9cbd36b01bd1cbaebf2de08d9173bc095c"; // WBNB
const approveAddress = "0x10ed43c718714eb63d5aa57b78b54704e256024e"; // PCRouter v2

const tokenContract = new ethers.Contract(
	tokenAddress,
	erc20ABI,
	provider(56).getSigner(address) // address - connected user address
);
const maxApproval = ethers.constants.MaxUint256;

const txRes = await tokenContract.approve(approveAddress, maxApproval, {
	gasLimit: 310000,
});
console.log(txRes);

}

Browser: Google Chrome desktop

fix: tighten `/metrics` to internal-only endpoint

We received a report that /metrics is open to the internet which it is.
Usually we expose this in a VPC private endpoint and should do the same here.

See https://walletconnect.slack.com/archives/C03SCF66K2T/p1676627585901279

feat: add websocket support

Add websocket connection support to RPC proxy

add `provider` to tracked rpc_requests data

as we want to know about how much rpc_requests are routed to which provider ( infura,quicknode,etc) we need to add a new field to the tracked data that contains the provider.

ac:

• rpc_requests logs have a provider column with the id/name of the provider

Set up alerts with Infura/Pokt

We should know when we reach out quotas. @chadyj is this something you could help us with?

Configure CORS

Monitor for Rate Limits on providers

e.g. Infura sends HTTP429 when throttled.

fix: latency and provider responses data missing in grafana

Provider responses by status code and latency data seems to be missing in grafana.
Will need to check grafana config/app to find the source of the problem

Move logging from error conversion functionality to proper axum layer

Nodar noticed that error logging is done in weird place. Proper way to do this (probably axum layer) needs research.
#115

feat: add scalling to rpc-proxy

Track integration latency

We should have a latency metric that subtracks the proxy time and only tracks the latency we are adding.

feat: add aurora support to rpc proxy

Aurora is gaining traction and is quite requested chain.
As it's provided by supported provider, shouldn't be a big change.

Integrate billing

fix: add logging to rpc proxy

Currently there's no way to figure out why RPC Proxy is throwing HTTP 5xx to the user, unless it's 502.
Logs are needed to be able to distinguish and understand the reason,