Add token usage and model tracking to Claude billing (#7729)

## Summary
- Adds token tracking fields to `misc.claude_code_usage` schema:
  - `input_tokens`, `output_tokens`
  - `cache_read_input_tokens`, `cache_creation_input_tokens`
  - `model`
- Updates `upload-claude-usage` action to extract these fields from
Claude output
- Updates S3 replicator lambda to handle new fields
- Creates v2 Grafana dashboard with token metrics

## Dashboard v2

https://pytorchci.grafana.net/public-dashboards/83058a8d65d44a099eb8d9ac2916f411

New metrics:
- Total input/output tokens
- Cache hit rate
- Cost by model breakdown
- Token usage by workflow
- Daily cache performance

## Test plan
- [ ] Deploy schema changes (ALTER TABLE to add columns)
- [ ] Deploy lambda changes
- [ ] Verify new data flows with token fields populated
- [ ] Verify v2 dashboard shows token metrics

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Opus 4.5 <[email protected]>

Author: wdvr

.github/actions/upload-claude-usage/action.yml

@@ -44,6 +44,11 @@ runs:
           --argjson duration_ms "$(echo "$RESULT" | jq -r '.duration_ms // 0')" \
           --argjson num_turns "$(echo "$RESULT" | jq -r '.num_turns // 0')" \
           --argjson total_cost_usd "$(echo "$RESULT" | jq -r '.total_cost_usd // 0')" \
+          --argjson input_tokens "$(echo "$RESULT" | jq -r '.usage.input_tokens // 0')" \
+          --argjson output_tokens "$(echo "$RESULT" | jq -r '.usage.output_tokens // 0')" \
+          --argjson cache_read_input_tokens "$(echo "$RESULT" | jq -r '.usage.cache_read_input_tokens // 0')" \
+          --argjson cache_creation_input_tokens "$(echo "$RESULT" | jq -r '.usage.cache_creation_input_tokens // 0')" \
+          --arg model "$(echo "$RESULT" | jq -r '.modelUsage | keys[0] // ""')" \
           '{
             repo: $repo,
             run_id: $run_id,
@@ -54,7 +59,12 @@ runs:
             timestamp: $timestamp,
             duration_ms: $duration_ms,
             num_turns: $num_turns,
-            total_cost_usd: $total_cost_usd
+            total_cost_usd: $total_cost_usd,
+            input_tokens: $input_tokens,
+            output_tokens: $output_tokens,
+            cache_read_input_tokens: $cache_read_input_tokens,
+            cache_creation_input_tokens: $cache_creation_input_tokens,
+            model: $model
           }' > /tmp/claude_usage.json
 
         aws s3 cp /tmp/claude_usage.json \

aws/lambda/clickhouse-replicator-s3/lambda_function.py

@@ -680,7 +680,12 @@ def claude_code_usage_adapter(table, bucket, key):
     `timestamp` DateTime64(3),
     `duration_ms` Int64,
     `num_turns` Int32,
-    `total_cost_usd` Float64
+    `total_cost_usd` Float64,
+    `input_tokens` Int64,
+    `output_tokens` Int64,
+    `cache_read_input_tokens` Int64,
+    `cache_creation_input_tokens` Int64,
+    `model` String
     """
     general_adapter(table, bucket, key, schema, ["none"], "JSONEachRow")
 

clickhouse_db_schema/misc.claude_code_usage/schema.sql

@@ -10,6 +10,11 @@ CREATE TABLE misc.claude_code_usage
     `duration_ms` Int64,
     `num_turns` Int32,
     `total_cost_usd` Float64,
+    `input_tokens` Int64 DEFAULT 0,
+    `output_tokens` Int64 DEFAULT 0,
+    `cache_read_input_tokens` Int64 DEFAULT 0,
+    `cache_creation_input_tokens` Int64 DEFAULT 0,
+    `model` String DEFAULT '',
     `_meta` Tuple(
         bucket String,
         key String),