Review and improve storage size limits documentation

[alongir]

Description

PR #78 adds a 2 GB per-snapshot size limit to three pages. This issue tracks reviewing whether all relevant pages document storage size limits well enough.

What PR #78 covers

• v2/traffic_snapshots.md — adds "Maximum size" row to constraints table
• dashboard_snapshots.md — adds "Maximum size" row to constraints table
• mcp/raw_capture_tools.md — adds callout note in POST /mcp/snapshots

Gaps to review

1. v2/raw_capture_config.md — The "Snapshot Storage" section documents storageSize for local PVCs but doesn't mention the 2 GB per-snapshot limit. Users configuring tap.snapshots.local.storageSize: 20Gi should understand that individual snapshots can't exceed 2 GB (even if the volume is larger). Add a note clarifying the distinction between total storage allocation and per-snapshot max.

2. snapshots_cloud_storage.md — No mention of the 2 GB limit. Users uploading to S3/Azure should know whether the limit applies to cloud uploads too (presumably yes, since it's a creation-time limit).

3. helm_reference.md — If there's a snapshots section, it should note the 2 GB per-snapshot limit alongside the storageSize config value.

4. Error behavior — None of the pages describe what happens when the limit is exceeded. Does snapshot creation fail? Is it truncated? What error does the user see? Documenting this helps users troubleshoot.

Why Is This Needed?

The 2 GB limit is a hard constraint that can surprise users who select a long time window or many nodes. Documenting it only in constraint tables may not be enough — users configuring storage via Helm or uploading to cloud need to see it in context.

Additional Context

• Related PR: Document 2GB snapshot size limitation #78
• The limit is per individual snapshot, not total storage
• Raw capture buffers themselves are not subject to this limit