diff --git a/config.toml b/config.toml
index 3d842fd..190db14 100644
--- a/config.toml
+++ b/config.toml
@@ -1,34 +1,39 @@
 # WeaselDB Configuration File
+# See config.md for complete documentation of all configuration options
 
 [server]
-# Network interfaces to listen on - production config with just TCP
+# Network interfaces where WeaselDB will accept connections
+# Options: TCP (address + port) or Unix domain sockets (path)
+# For production, use TCP. For local testing, consider Unix sockets for better performance
 interfaces = [
   { type = "tcp", address = "127.0.0.1", port = 8080 }
 ]
-# Maximum request size in bytes (for 413 Content Too Large responses)
+# Maximum size allowed for incoming requests (larger requests are rejected)
+# Increase if you need to handle very large transaction payloads
 max_request_size_bytes = 1048576  # 1MB
-# Number of I/O threads for handling connections and network events
+# Number of worker threads handling network connections
+# Start with 1, increase if CPU usage is high under load
 io_threads = 1
-# Event batch size for epoll processing
+# Internal network processing batch size
+# Higher values may improve throughput at cost of latency
 event_batch_size = 32
 
 [commit]
-# Minimum length for request_id to ensure sufficient entropy
+# Required minimum length for transaction request IDs
+# Longer IDs reduce chance of accidental duplicates across clients
 min_request_id_length = 20
-# How long to retain request IDs for /v1/status queries (hours)
+# How long to keep transaction status information available (hours)
+# Used by status API to look up the outcome of completed transactions
 request_id_retention_hours = 24
-# Minimum number of versions to retain request IDs
+# Alternative retention policy: keep transaction status for at least this many database versions
+# Ensures status lookups work even during periods of low database activity
 request_id_retention_versions = 100000000
 
 [subscription]
-# Maximum buffer size for unconsumed data in /v1/subscribe (bytes)
+# Memory limit for buffering change stream data per subscriber (bytes)
+# Subscribers that fall behind will be disconnected when this limit is reached
+# See api.md for details on the subscription streaming API
 max_buffer_size_bytes = 10485760  # 10MB
-# Interval for sending keepalive comments to prevent idle timeouts (seconds)
+# How often to send keep-alive messages to streaming subscribers (seconds)
+# Prevents network timeouts during periods of no database activity
 keepalive_interval_seconds = 30
-
-[benchmark]
-# CPU-intensive loop iterations for /ok requests in resolve stage
-# 0 = health check only (no CPU work)
-# 4000 = default benchmark load (740ns CPU work, 1M req/s)
-# Higher values = more CPU stress testing
-ok_resolve_iterations = 0