@FluxRPC Or self-host your own for free 👀 Built a Solana priority fee oracle in Go: 438k RPS, p50 157µs, zero API keys needed https://t.co/JqfH9UO8VI MIT license. Run it next to your RPC node 🔧
@Helius Great improvements! For teams that need even lower latency, self-hosted gas oracle next to your own RPC node is worth considering — I built one in Go: https://t.co/0xRQh0u4ih 438k RPS, p50 157µs. Happy to get feedback from the Helius team.
@mert Great improvements! For teams that need even lower latency, self-hosted gas oracle next to your own RPC node is worth considering — I built one in Go: https://t.co/0xRQh0u4ih 438k RPS, p50 157µs. Happy to get feedback from the Helius team.
Fixed bugs in Solana Gas Oracle:
- nil stop channel in APIHandler causing panic on shutdown
- MEV cache initialization with atomic.Pointer
- buffer pool type assertions and capacity validation
https://t.co/oR83B9Bukf
Added Docker support 🐳
#Solana#MEV#Blockchain
Next steps for the Solana Gas Oracle:
Docker + Kubernetes
Multi-RPC support
Jito integration
Better test coverage
Open source:
https://t.co/oR83B9C29N
#Solana#Golang#MEV#LowLatency#OpenSource
@Circular_fi@solana@orca_so@Raydium@Pumpfun Built this for exactly this type of workload.
Open-source Solana gas oracle:
~440k RPS
~157µs p50 latency
https://t.co/oR83B9Bukf
Most gas APIs are network-bound.
This Solana gas oracle runs fully in memory:
~440k requests/sec
~157µs p50 latency
Open source (MIT):
https://t.co/oR83B9Bukf
#solana#golang#mev
Built a Solana gas oracle for low-latency execution.
~440k RPS | ~157µs p50
In-memory design:
- ring buffers
- atomic snapshots
- O(n) QuickSelect percentiles
- minimal locking, no DB
Open source (MIT):
https://t.co/oR83B9C29N
Feedback from Solana infra builders welcome.
Still relying on lagging public APIs for Solana gas?
Check the benchmark of my self-hosted Go Oracle.
⚡️ Direct WebSocket mempool streaming.
⚡️ Zero third-party dependencies.
Landing transactions in 2026 requires microsecond precision. DM for the source code. 🛠️ #SolanaDev#MEV
Built a high-performance Solana gas oracle.
~438k RPS | ~157µs p50 | ~4.3ms p99
In-memory pipeline, no DB, no queues.
Go + fasthttp + lock-minimized design.
MIT License
GitHub:
https://t.co/oR83B9C29N
#solana#golang#backend#lowlatency
• Hyper-V virtualized environment
• HT enabled
• No CPU pinning / isolation
• Solana test validator running in background
This is a noisy setup by design.
Stable p99 under sustained load is what matters here.
#Solana#Infrastructure
Sustained load test (400s) under real conditions.
• 100k RPS sustained
• p50: 169µs
• p99: ~5ms
• No latency drift
• No errors
Dockerized service + live Solana validator running.
This is not a micro-benchmark.
#Solana#Infrastructure
410k+ RPS. 158µs Median Latency. 🦀⚡️
Pure Golang. Zero-overhead pipeline.
The ultimate Solana edge:
🔹 158µs median response
🔹 4.1M requests in 10s
🔹 278MB/s throughput
Source + IP exclusivity.
📥 DM for inquiries.
#Solana#MEV#Golang#Infrastructure
For people tired of vaporware:
✅ live binary
✅ real wrk / wrk2 benchmarks
✅ htop under load
✅ /stats endpoint with live metrics
No “soon™”. Everything is reproducible.
#solana#mev#infra#web3#backend
410k+ RPS. 158µs Median Latency. 🦀⚡️
Pure Golang. Zero-overhead pipeline.
The ultimate Solana edge:
🔹 158µs median response
🔹 4.1M requests in 10s
🔹 278MB/s throughput
Source + IP exclusivity.
📥 DM for inquiries.
#Solana#MEV#Golang#Infrastructure
🚀 Solana MEV / Gas API — live, not a concept
Built a production-ready service that:
• streams real-time priority fees
• handles ~400k req/sec
• p50 < 200µs, p99 < 5ms
• built-in rate limiting + caching
• zero hype, just code
Benchmarked on a Hyper-V virtualized environment (hence the P99 jitter). On bare metal, expect these numbers to compress significantly.
Tested with wrk2 to ensure consistent throughput under extreme pressure.
Real-world performance (at 10-50k RPS) stays at near-median latency.