- Result: 247/329 working (75% success rate)
- Test: Connection + version call + transaction fetch
- Timeout: 5 seconds per server
- Duration: ~10 minutes for full test
Top 30 Fastest Servers:
- guichet.centure.cc - 58ms (TCP)
- guichet.centure.cc - 87ms (SSL)
- electrum.loyce.club - 91ms (TCP)
- horsey.cryptocowboys.net - 95ms (TCP)
- 52.1.56.181 - 110ms (TCP)
Statistics:
- Average latency: 1.182s
- SSL servers: 127
- TCP servers: 120
- Fastest: guichet.centure.cc (58ms)
- Slowest: ecdsa.net (2.789s)
Backend:
- Single Electrum server mode (working)
- 247 verified servers in fallback list
- API endpoint:
/api/all-serversreturns full test results
Frontend:
- Main app: http://localhost:5173/
- Server list: http://localhost:5173/servers (247 servers with real test data)
- Index: http://localhost:5173/index
Multiplexer (DISABLED):
- Initializing 50+ connections blocks startup (hangs)
- Many servers from 1209k.com are dead/unreachable (DNS errors, timeouts)
- Pool of 15-50 connections takes too long to initialize
Root Cause: Most servers from 1209k.com are garbage. Even with 247 "working" servers, only a handful are actually reliable.
Keep:
backend/working_servers.json- Full test results (247 servers)backend/app/services/electrum_servers.py- Updated with 247 verified serversfrontend/src/pages/ServerListPage.tsx- Simple table showing all test results
Infrastructure (keep but unused):
backend/app/services/electrum_pool.py- Connection pool (disabled)backend/app/services/electrum_multiplexer.py- High-level API (disabled)backend/app/api/metrics.py- Metrics endpointsbackend/app/api/electrum_proxy.py- Electrum proxy endpoint
Use single-server mode with top 5 fastest servers as fallbacks:
- guichet.centure.cc:50001 (TCP) - 58ms
- electrum.loyce.club:50001 (TCP) - 91ms
- horsey.cryptocowboys.net:50001 (TCP) - 95ms
- fulcrum1.getsrt.net:50002 (SSL) - 126ms
- smmalis37.ddns.net:50001 (TCP) - 141ms
The multiplexer adds complexity without benefit when most servers are unreliable.