[Feature] Add weighted load balancing support for RPC providers
@DaMandal0rian já está trabalhando nisso.
Desde 14/8/2025.
Avaliação
Esta issue ainda não foi avaliada.
Descrição
Problem
Currently, Graph Node uses a sequential failover approach for RPC providers. It attempts to use the first configured provider and only moves to the next provider if there are issues or if a required feature is not available. This means traffic is not distributed across providers - instead, the first healthy provider handles all requests until it fails.
This limitation prevents operators from:
- Distributing load across multiple healthy RPC endpoints
- Utilizing all available RPC capacity efficiently
- Preventing single provider overload when multiple endpoints are available
- Optimizing costs by balancing requests across providers with different pricing
- Achieving true high availability through active load distribution
Proposed Solution
Implement weighted random load balancing for RPC provider selection, allowing operators to configure traffic distribution percentages for each endpoint.
Example Use Cases
- Performance-based routing: Direct 80% of traffic to a high-performance dedicated node and 20% to a backup
- Cost optimization: Route 90% to cheaper endpoints and 10% to premium endpoints for critical requests
- Gradual rollout: Start a new provider at 5% traffic and increase as confidence grows
- Capacity management: Distribute load according to each endpoint's actual capacity
Implementation
I've created PR #[YOUR_PR_NUMBER] that adds this functionality with:
- Configurable decimal weights (0.0-1.0) per RPC provider
- New
--weighted-rpc-steeringCLI flag and config option - Backward compatibility (defaults to equal weights when not specified)
- Weighted random selection algorithm for provider choice
Configuration Example
[chains.mainnet]
weighted_rpc_steering = true
provider = [
{
label = "primary-node",
details = { type = "web3", url = "https://fast.example.com", weight = 0.7 }
},
{
label = "secondary-node",
details = { type = "web3", url = "https://medium.example.com", weight = 0.2 }
},
{
label = "backup-node",
details = { type = "web3", url = "https://backup.example.com", weight = 0.1 }
}
]
Related Work
- PR - Implementation of weighted load balancing
- This addresses a common request for managing multiple RPC endpoints with varying characteristics
Are you aware of any blockers that must be resolved before implementing this feature? If so, which? Link to any relevant GitHub issues.
No response
Some information to help us out
- Tick this box if you plan on implementing this feature yourself.
- I have searched the issue tracker to make sure this issue is not a duplicate.
- Linguagem predominante
- Rust
- Estrelas
- 3.2k
- Forks
- 1.1k
- Métricas de merge de PRs
- Nenhum PR com merge em 30d
Preparar o ambiente
Inicia o contêiner de desenvolvimento do projeto no navegador, com a sua própria conta do GitHub.
- Sem Dockerfile nem arquivo Docker Compose
- Tem um modelo de pull request
- Ler o guia de contribuição
Primeiros passos
- Leia a issue inteira e depois o guia de contribuição do projeto.
- Comente na issue dizendo que vai assumir — evita que duas pessoas façam o mesmo trabalho.
- Faça um fork do repositório e trabalhe em uma branch.
- Abra um pull request que referencie o número da issue.
Mais de graphprotocol/graph-node
-
current: include emits an all-null bucket for dimensionless aggregations, nulling the whole responseAberta
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 78/100
graphprotocol/graph-node#6719 ·
-
RUSTSEC-2026-0194: Quadratic run time when checking a start tag for duplicate attribute namesTalvez já em andamento @szupzj18 assumiu há 58 dias. Aberta
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 68/100
graphprotocol/graph-node#6673 ·
-
RUSTSEC-2026-0185: Remote memory exhaustion in quinn-proto from unbounded out-of-order stream reassemblyTalvez já em andamento @abisheik687 assumiu há 99 dias. Aberta
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 70/100
graphprotocol/graph-node#6650 · 1 comentário ·
-
`loadRelated` can return stale children while blocks are queued for writing (`!= any` in `FindDerivedQuery`; unrelated queued writes not excluded)Talvez já em andamento @madumas assumiu há 5 dias. Aberta
Dificuldade 4/5 3-5 dias Facilidade para iniciantes 50/100
graphprotocol/graph-node#6726 ·
-
Dificuldade 4/5 3-5 dias Facilidade para iniciantes 48/100
graphprotocol/graph-node#6722 ·
Todas as issues de graphprotocol/graph-node
Issues semelhantes
-
✨ enhancement needs-discussion
Dificuldade 1/5 Menos de uma hora Facilidade para iniciantes 85/100
-
docs(openclaw): RTK_REWRITE_HOST relaxes every default ask, not only commands no rule matchedAbertaarea:docs documentation good first issue priority:low
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 68/100
rtk-ai/rtk#4500 · 1 comentário ·
Mantenedores costumam responder em até 1 dia
-
triage:accepted
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 65/100
open-telemetry/otel-arrow#4343 ·
Mantenedores costumam responder em até 2 dias
-
bug
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 75/100
mishraprafful/multihull#150 ·
Mantenedores costumam responder em até 1 dia
-
area:tooling bug good first issue priority:P3
Dificuldade 2/5 1-3 horas Facilidade para iniciantes 72/100
michaelnavazhylau/ngspice-rs#129 ·
Mantenedores costumam responder em até 1 dia