GitChain Edge Cache Management

Manage the Cloudflare KV cache that powers GitChain's edge search.

Process

1. Parse $ARGUMENTS for action:
   • warm [namespace] — Pre-populate cache with all products from a namespace
   • purge [pattern] — Clear cache entries matching pattern
   • stats — Show cache hit/miss rates and entry counts
   • inspect [key] — Show cached value for a specific key
2. Use Cloudflare MCP for KV operations
3. Use GitChain search/get for warming

Cache Warming

For warm bosch:

1. GitChain search({namespace: "bosch", limit: 100}) — paginate through all
2. For each product: get({id}) → extract key fields → write to KV
3. KV key format: product:{namespace}:{identifier}
4. KV value: compact JSON (name, article_nr, key_specs, price, dimensions)
5. TTL: 86400 (24 hours)

Output Format

# Cache Operation: [action]

## Results
| Metric | Value |
|--------|-------|
| Entries written | X |
| Entries purged | X |
| Total cache size | X entries |
| Estimated KV cost | $X/month |

## Cache Contents by Namespace
| Namespace | Entries | Avg Size | TTL |
|-----------|---------|----------|-----|
| bosch | X | X KB | 24h |
| lightnet | X | X KB | 24h |

Rules

• Warm in batches of 50 to avoid rate limits
• Never cache more than 10,000 entries (KV free tier: 100K reads/day)
• Always show cost estimate
• Include TTL in all writes