The live route tables lose their random UUID key; the row id the
prefix-length sample draws is declared as the integer key. The prefix
tables are keyed by prefix and keep only first_seen; a batch inserts only
prefixes not stored yet. A route update sets only the columns that
change. Each kept index names the queries using it. Every write goes
through one connection kept for writing, with a 1 GiB page cache; reading
connections get 16 MiB, heap limits rise to 1.5 and 2 GiB (README
"Memory"). Batch writes go in key order. A benchmark run by hand writes
250,000 updates on 5 million routes faster than the feed's 10,000 a
second; a live run follows.
Model: opus-5-5
Takes .golangci.yml, .dockerignore, .gitignore, .editorconfig, the
workflow and REPO_POLICIES.md from sneak/prompts 0653579, with this
repository's own entries after the shared content. Lint (golangci-lint
v2.14.0) and test are Dockerfile phases that script/lint and script/test
build alone with a new CACHEBUST value, keeping Go's and golangci-lint's
caches; the image build depends on both. make fmt and make fmt-check also
format the markdown with prettier, pinned in yarn.lock. The agent
guidance moves to AGENTS.md. The Go code is fixed for what the new
configuration finds, and the AS and prefix detail reads now return the
rows' error after their loops.
Model: opus-5-5
- Add context-aware versions of all read operations in the database
- Update handlers to use context from HTTP requests
- Allows database queries to be cancelled when HTTP requests timeout
- Prevents database operations from continuing after client disconnects
- Create internal/logger package with Logger wrapper around slog
- Logger automatically adds source file, line number, and function name to all log entries
- Use golang.org/x/term to properly detect if stdout is a terminal
- Replace all slog.Logger usage with logger.Logger throughout the codebase
- Remove verbose logging from database GetStats() method
- Update all constructors and dependencies to use the new logger
- Remove live_routes table from SQL schema and all related indexes
- Create new internal/routingtable package with thread-safe RoutingTable
- Implement RouteKey-based indexing with secondary indexes for efficient lookups
- Add RoutingTableHandler to manage in-memory routes separately from database
- Update DatabaseHandler to only handle persistent database operations
- Wire up RoutingTable through fx dependency injection
- Update server to get live route count from routing table instead of database
- Remove LiveRoutes field from database.Stats struct
- Update tests to work with new architecture
- Set concurrent handlers back to 100 (from 200)
- Set slow query threshold to 50ms (from 10ms)
These values provide a good balance between throughput and
system resource usage.
Queries in the 50-70ms range are acceptable for now given SQLite's
write serialization constraints. Setting threshold to 100ms to focus
on truly problematic queries.
The 10ms threshold was too noisy - queries in the 10-20ms range are
actually performing well after the index optimizations. Setting the
threshold to 50ms will help identify truly problematic queries.
- Implement comprehensive SQL query logging for queries over 10ms
- Add logging wrapper methods for all database operations
- Replace timing code in GetStats with simple info log messages
- Add missing database indexes for better query performance:
- idx_live_routes_lookup for common prefix/origin/peer lookups
- idx_live_routes_withdraw for withdrawal updates
- idx_prefixes_prefix for prefix lookups
- idx_asn_peerings_lookup for peering relationship queries
- Increase SQLite cache size to 512MB
- Add performance-oriented SQLite pragmas
- Extract HTML templates to separate files using go:embed
- Add JSON response middleware with @meta field (like bgpview.io API)
- Fix concurrent map write errors in HTTP handlers
- Add request timeout handling with proper JSON error responses
These changes significantly improve database query performance and
provide visibility into slow queries for debugging purposes.