A reverse proxy / load balancer for axon.

Features

• HTTP and WebSocket. Subscription is supported on WebSocket connections.

• P2C least requests or client ip consistent hashing load balancing.

• Filter id sticky load balancing.

  E.g. an eth_getFilterChanges request will be forwarded to the node that creates the filter.

• Health check and failover.

• Per-ip (and per-ip-per-method) JSONRPC rate limiting with redis.

• JSONRPC caching with redis.

• High availability: running multiple instances of this proxy is supported.

• Metrics.

Not Planned

Features like https, http/2, http/3 or more complicated CORS configurations. These should be supported by a dedicated http proxy in front of axon-proxy if necessary.

Building

Install rust and

$ cargo build --release

A dockerfile is also provided.

Running

$ axon-proxy -c config.toml

Specify the number of worker threads with --threads or AXON_PROXY_THREADS:

$ axon-proxy --threads 8 -c config.toml

The default is to use the same number of threads as the number of CPU cores.

Configuration

bind = 8000
redis.url = "redis://127.0.0.1/"
# Upstream WebSocket connections are only used for subscription. Random LB is used for now.
ws_nodes = ["ws://127.0.0.1:3001/ws", "ws://127.0.0.2:3001/ws"]
# Default is p2c_least_requests
# lb = "client_ip_hashing"

[[nodes]]
url = "http://127.0.0.1:3000/"

[[nodes]]
url = "http://127.0.0.2:3000/"
# Weight is only effective if lb is client_ip_hashing.
weight = 0.5

# Health check is enabled by default.
# [health_check]
# enabled = true
# timeout_secs = 3
# interval_secs = 3
# concurrency = 8

# Default is no caching.
[cache]
eth_call = true
eth_estimateGas = true

# Default is no rate limit.
# [rate_limit]
# total = 3000
# method."eth_call" = 600

# Override rate limit settings for specific client ips.
# Method rate limits are NOT inherited.
# [rate_limit.ip."127.0.0.1"]
# total = 10000
# method."eth_call" = 2000