Recording Limits

Recording Limits let you bound how much traffic Madhyamas records to prevent unbounded memory and disk usage during long debugging sessions. When limits are exceeded, the oldest entries are automatically pruned (FIFO — first in, first out).

Why Recording Limits Matter

Without limits, a long-running proxy session capturing a high-traffic app can accumulate gigabytes of data, leading to:

• Slow web UI performance (large database queries)
• High memory usage
• Disk space exhaustion
• Degraded proxy throughput

Recording limits keep Madhyamas responsive regardless of traffic volume.

Configuration Options

All options can be set via the web UI (Config → Capture tab), the REST API, or the config file. Changes take effect immediately — no restart required.

Option	Default	Description
Max Requests	10,000	Maximum number of traffic entries to keep. When exceeded, oldest entries are pruned.
Max Total Size	Unlimited	Maximum total size of all stored bodies in MB. When exceeded, oldest entries are pruned. Set to empty/null to disable.
Max Body Size	20 MB	Maximum size of a single request or response body. Larger bodies are truncated before storage.
Capture Request Bodies	true	Whether to store request bodies. When off, request headers and metadata are still recorded.
Capture Response Bodies	true	Whether to store response bodies. When off, response headers and metadata are still recorded.
Ignored Domains	(empty)	Domains whose traffic should not be recorded at all. Supports exact match, suffix match, and wildcard patterns.

How Pruning Works

Pruning uses a FIFO (first in, first out) strategy — the oldest entries (by timestamp) are deleted first.

• Entry count limit (max_requests): checked on every insert. If the count exceeds the limit, the surplus oldest entries are pruned immediately. Set to 0 to disable.
• Total size limit (max_total_size_mb): checked periodically. If the total body size exceeds the limit, oldest entries are pruned one by one until under the limit. Set to null to disable.

When entries are pruned, the web UI removes them from the traffic list in real time — no manual refresh needed.

Ignored Domains

The ignored_domains setting prevents traffic from specified hosts from being recorded at all. This is useful for filtering out noise from analytics, telemetry, or CDN domains.

Matching is case-insensitive and supports three patterns:

Pattern	Example	Matches
Exact hostname	example.com	example.com
Suffix match	example.com	example.com, api.example.com, www.example.com
Wildcard subdomain	*.example.com	api.example.com, www.example.com (not example.com itself)

Example:

*.google-analytics.com
*.doubleclick.net
telemetry.example.com

Recording Quota Indicator

The web UI header displays a recording quota indicator showing the current entry count versus the limit (e.g. 1,234/10,000). The indicator turns amber when usage exceeds 80% of the limit. Hovering over it shows the total recording size versus the size limit (when configured).

From the Web UI

1. Click the Config button in the header
2. Navigate to the Capture tab
3. Adjust the settings:
   • Capture Request Bodies — toggle on/off
   • Capture Response Bodies — toggle on/off
   • Max Body Size — slider
   • Max Total Recording Size — input in MB (empty = unlimited)
   • Ignored Domains — textarea, one pattern per line
4. Click Save Changes

From the CLI / API

# View current stats
curl http://127.0.0.1:3001/api/capture/stats

# Set max entries to 5,000 and total size to 500 MB
curl -X PATCH http://127.0.0.1:3001/api/config \
  -H "Content-Type: application/json" \
  -d '{"max_requests": 5000, "max_total_size_mb": 500}'

# Disable response body capture
curl -X PATCH http://127.0.0.1:3001/api/config \
  -H "Content-Type: application/json" \
  -d '{"capture_response_bodies": false}'

# Add ignored domains
curl -X PATCH http://127.0.0.1:3001/api/config \
  -H "Content-Type: application/json" \
  -d '{"ignored_domains": ["*.google-analytics.com", "telemetry.example.com"]}'

Common Use Cases

Long-Running Captures

Set a max_total_size_mb so an overnight capture can't fill your disk — old entries are pruned automatically as new traffic arrives.

Reducing Noise

Add analytics and telemetry domains to ignored_domains so they never appear in your traffic list, keeping it focused on the requests you actually care about.

Saving Memory on Low-Power Machines

Lower max_requests and disable request body capture to reduce memory and disk usage on laptops or VMs with limited resources.

Body-Only Debugging

If you only care about response bodies (e.g. debugging API payloads), disable request body capture to halve the storage overhead while keeping full response inspection.

See also

• Configuration — max_requests, max_body_size, and related settings
• Auto Save — periodic backups of the session
• Performance — memory tracking and tuning
• REST API reference — /api/capture/stats