feat: track step runtime_ms on chat messages (#23219)

## Summary

Adds a `runtime_ms` column to `chat_messages` that records the
wall-clock duration (in milliseconds) of each LLM step. This covers LLM
streaming, tool execution, and retries — the full time the agent is
"alive" for a step.

This is the foundation for billing by agent alive time. The column
follows the same pattern as `total_cost_micros`: stored per assistant
message, aggregatable with `SUM()` over time periods by user.

## Changes

- **Migration**: adds nullable `runtime_ms bigint` to `chat_messages`.
- **chatloop**: adds `Runtime time.Duration` field to `PersistedStep`,
measures `time.Since(stepStart)` at the beginning of each step (covering
stream + tool execution + retries).
- **chatd**: passes `step.Runtime.Milliseconds()` to the assistant
message `InsertChatMessage` call; all other message types (system, user,
tool) get `NULL`.
- **Tests**: adds `runtime > 0` assertion in chatloop tests.

## Billing query pattern

Once ready, aggregation mirrors the existing cost queries:

```sql
SELECT COALESCE(SUM(cm.runtime_ms), 0)::bigint AS total_runtime_ms
FROM chat_messages cm
JOIN chats c ON c.id = cm.chat_id
WHERE c.owner_id = @user_id
  AND cm.created_at >= @start_time
  AND cm.created_at < @end_time
  AND cm.runtime_ms IS NOT NULL;
```
This commit is contained in:
Kyle Carberry
2026-03-18 10:57:35 -04:00
committed by GitHub
parent 3bcb7de7c0
commit 4dd8531f37
10 changed files with 59 additions and 12 deletions
+7
View File
@@ -42,6 +42,11 @@ type PersistedStep struct {
Content []fantasy.Content
Usage fantasy.Usage
ContextLimit sql.NullInt64
// Runtime is the wall-clock duration of this step,
// covering LLM streaming, tool execution, and retries.
// Zero indicates the duration was not measured (e.g.
// interrupted steps).
Runtime time.Duration
}
// RunOptions configures a single streaming chat loop run.
@@ -260,6 +265,7 @@ func Run(ctx context.Context, opts RunOptions) error {
for step := 0; totalSteps < opts.MaxSteps; step++ {
totalSteps++
stepStart := time.Now()
// Copy messages so that provider-specific caching
// mutations don't leak back to the caller's slice.
// copy copies Message structs by value, so field
@@ -365,6 +371,7 @@ func Run(ctx context.Context, opts RunOptions) error {
Content: result.content,
Usage: result.usage,
ContextLimit: contextLimit,
Runtime: time.Since(stepStart),
}); err != nil {
if errors.Is(err, ErrInterrupted) {
persistInterruptedStep(ctx, opts, &result)
+3
View File
@@ -7,6 +7,7 @@ import (
"strings"
"sync"
"testing"
"time"
"charm.land/fantasy"
fantasyanthropic "charm.land/fantasy/providers/anthropic"
@@ -64,6 +65,8 @@ func TestRun_ActiveToolsPrepareBehavior(t *testing.T) {
require.Equal(t, 1, persistStepCalls)
require.True(t, persistedStep.ContextLimit.Valid)
require.Equal(t, int64(4096), persistedStep.ContextLimit.Int64)
require.Greater(t, persistedStep.Runtime, time.Duration(0),
"step runtime should be positive")
require.NotEmpty(t, capturedCall.Prompt)
require.False(t, containsPromptSentinel(capturedCall.Prompt))