shishantbiswas/coder
1
0
Fork 0
mirror of https://github.com/coder/coder.git synced 2026-06-02 20:48:20 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat: configure acquire chat batch size (#23196)

Browse Source 
## Summary
- add a hidden deployment config option for chat acquire batch size
(`CODER_CHAT_ACQUIRE_BATCH_SIZE` / `chat.acquireBatchSize`)
- thread the configured value into chatd startup while preserving the
existing default of `10`
- clamp the deployment value to the `int32` range before passing it into
chatd
- regenerate the API/docs/types/testdata artifacts for the new config
field

## Why
`chatd` currently acquires pending chats in batches of `10` via a
compile-time default. This change makes that batch size
operator-configurable from deployment config, so we can tune acquisition
behavior without another code change.
This commit is contained in:
Ethan
2026-03-19 00:54:32 +11:00
committed by GitHub
parent 8b4d35798a
commit fc3508dc60
9 changed files with 113 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files
cli/testdata/server-config.yaml.golden
Vendored
+5
View File
@@ -752,6 +752,11 @@ workspace_prebuilds:
# limit; disabled when set to zero.
# (default: 3, type: int)
failure_hard_limit: 3
# Configure the background chat processing daemon.
chat:
# How many pending chats a worker should acquire per polling cycle.
# (default: 10, type: int)
acquireBatchSize: 10
aibridge:
# Whether to start an in-memory aibridged instance.
# (default: false, type: bool)
coderd/apidoc/docs.go
Generated
+11
View File
@@ -12714,6 +12714,9 @@ const docTemplate = `{
},
"bridge": {
"$ref": "#/definitions/codersdk.AIBridgeConfig"
},
"chat": {
"$ref": "#/definitions/codersdk.ChatConfig"
}
}
},
@@ -13771,6 +13774,14 @@ const docTemplate = `{
}
}
},
"codersdk.ChatConfig": {
"type": "object",
"properties": {
"acquire_batch_size": {
"type": "integer"
}
}
},
"codersdk.ConnectionLatency": {
"type": "object",
"properties": {
coderd/apidoc/swagger.json
Generated
+11
View File
@@ -11320,6 +11320,9 @@
},
"bridge": {
"$ref": "#/definitions/codersdk.AIBridgeConfig"
},
"chat": {
"$ref": "#/definitions/codersdk.ChatConfig"
}
}
},
@@ -12342,6 +12345,14 @@
}
}
},
"codersdk.ChatConfig": {
"type": "object",
"properties": {
"acquire_batch_size": {
"type": "integer"
}
}
},
"codersdk.ConnectionLatency": {
"type": "object",
"properties": {
coderd/chatd/chatd.go
+11 -3
View File
@@ -58,11 +58,11 @@ const (
// of 5 means recovery runs at 1/5 of the stale-after duration.
staleRecoveryIntervalDivisor = 5
// maxChatsPerAcquire is the maximum number of chats to
// DefaultMaxChatsPerAcquire is the maximum number of chats to
// acquire in a single processOnce call. Batching avoids
// waiting a full polling interval between acquisitions
// when many chats are pending.
maxChatsPerAcquire int32 = 10
DefaultMaxChatsPerAcquire int32 = 10
defaultSubagentInstruction = "You are running as a delegated sub-agent chat. Complete the delegated task and provide clear, concise assistant responses for the parent agent."
)
@@ -98,6 +98,7 @@ type Server struct {
// Configuration
pendingChatAcquireInterval time.Duration
maxChatsPerAcquire int32
inFlightChatStaleAfter time.Duration
}
@@ -1174,6 +1175,7 @@ type Config struct {
ReplicaID uuid.UUID
SubscribeFn SubscribeFn
PendingChatAcquireInterval time.Duration
MaxChatsPerAcquire int32
InFlightChatStaleAfter time.Duration
AgentConn AgentConnFunc
CreateWorkspace chattool.CreateWorkspaceFn
@@ -1199,6 +1201,11 @@ func New(cfg Config) *Server {
inFlightChatStaleAfter = DefaultInFlightChatStaleAfter
}
maxChatsPerAcquire := cfg.MaxChatsPerAcquire
if maxChatsPerAcquire <= 0 {
maxChatsPerAcquire = DefaultMaxChatsPerAcquire
}
workerID := cfg.ReplicaID
if workerID == uuid.Nil {
workerID = uuid.New()
@@ -1219,6 +1226,7 @@ func New(cfg Config) *Server {
providerAPIKeys: cfg.ProviderAPIKeys,
instructionCache: make(map[uuid.UUID]cachedInstruction),
pendingChatAcquireInterval: pendingChatAcquireInterval,
maxChatsPerAcquire: maxChatsPerAcquire,
inFlightChatStaleAfter: inFlightChatStaleAfter,
}
@@ -1272,7 +1280,7 @@ func (p *Server) processOnce(ctx context.Context) {
chats, err := p.db.AcquireChats(acquireCtx, database.AcquireChatsParams{
StartedAt: time.Now(),
WorkerID: p.workerID,
NumChats: maxChatsPerAcquire,
NumChats: p.maxChatsPerAcquire,
})
acquireCancel()
if err != nil {
coderd/coderd.go
+20 -10
View File
@@ -10,6 +10,7 @@ import (
"flag"
"fmt"
"io"
"math"
"net/http"
httppprof "net/http/pprof"
"net/url"
@@ -766,17 +767,26 @@ func New(options *Options) *API {
}
api.agentProvider = stn
maxChatsPerAcquire := options.DeploymentValues.AI.Chat.AcquireBatchSize.Value()
if maxChatsPerAcquire > math.MaxInt32 {
maxChatsPerAcquire = math.MaxInt32
}
if maxChatsPerAcquire < math.MinInt32 {
maxChatsPerAcquire = math.MinInt32
}
api.chatDaemon = chatd.New(chatd.Config{
Logger: options.Logger.Named("chats"),
Database: options.Database,
ReplicaID: api.ID,
SubscribeFn: options.ChatSubscribeFn,
ProviderAPIKeys: chatProviderAPIKeysFromDeploymentValues(options.DeploymentValues),
AgentConn: api.agentProvider.AgentConn,
CreateWorkspace: api.chatCreateWorkspace,
StartWorkspace: api.chatStartWorkspace,
Pubsub: options.Pubsub,
WebpushDispatcher: options.WebPushDispatcher,
Logger: options.Logger.Named("chats"),
Database: options.Database,
ReplicaID: api.ID,
SubscribeFn: options.ChatSubscribeFn,
MaxChatsPerAcquire: int32(maxChatsPerAcquire), //nolint:gosec // maxChatsPerAcquire is clamped to int32 range above.
ProviderAPIKeys: chatProviderAPIKeysFromDeploymentValues(options.DeploymentValues),
AgentConn: api.agentProvider.AgentConn,
CreateWorkspace: api.chatCreateWorkspace,
StartWorkspace: api.chatStartWorkspace,
Pubsub: options.Pubsub,
WebpushDispatcher: options.WebPushDispatcher,
})
gitSyncLogger := options.Logger.Named("gitsync")
refresher := gitsync.NewRefresher(
codersdk/deployment.go
+22
View File
@@ -1437,6 +1437,11 @@ func (c *DeploymentValues) Options() serpent.OptionSet {
Parent: &deploymentGroupNotifications,
YAML: "inbox",
}
deploymentGroupChat = serpent.Group{
Name: "Chat",
YAML: "chat",
Description: "Configure the background chat processing daemon.",
}
deploymentGroupAIBridge = serpent.Group{
Name: "AI Bridge",
YAML: "aibridge",
@@ -3600,6 +3605,18 @@ Write out the current server config as YAML to stdout.`,
Group: &deploymentGroupClient,
YAML: "hideAITasks",
},
// Chat Options
{
Name: "Chat: Acquire Batch Size",
Description: "How many pending chats a worker should acquire per polling cycle.",
Flag: "chat-acquire-batch-size",
Env: "CODER_CHAT_ACQUIRE_BATCH_SIZE",
Value: &c.AI.Chat.AcquireBatchSize,
Default: "10",
Group: &deploymentGroupChat,
YAML: "acquireBatchSize",
Hidden: true, // Hidden because most operators should not need to modify this.
},
// AI Bridge Options
{
Name: "AI Bridge Enabled",
@@ -4052,9 +4069,14 @@ type AIBridgeProxyConfig struct {
UpstreamProxyCA serpent.String `json:"upstream_proxy_ca" typescript:",notnull"`
}
type ChatConfig struct {
AcquireBatchSize serpent.Int64 `json:"acquire_batch_size" typescript:",notnull"`
}
type AIConfig struct {
BridgeConfig AIBridgeConfig `json:"bridge,omitempty"`
BridgeProxyConfig AIBridgeProxyConfig `json:"aibridge_proxy,omitempty"`
Chat ChatConfig `json:"chat,omitempty" typescript:",notnull"`
}
type SupportConfig struct {
docs/reference/api/general.md
Generated
+3
View File
@@ -204,6 +204,9 @@ curl -X GET http://coder-server:8080/api/v2/deployment/config \
"retention": 0,
"send_actor_headers": true,
"structured_logging": true
},
"chat": {
"acquire_batch_size": 0
}
},
"allow_workspace_renames": true,
docs/reference/api/schemas.md
Generated
+24
View File
@@ -786,6 +786,9 @@
"retention": 0,
"send_actor_headers": true,
"structured_logging": true
},
"chat": {
"acquire_batch_size": 0
}
}
```
@@ -796,6 +799,7 @@
|------------------|--------------------------------------------------------------|----------|--------------|-------------|
| `aibridge_proxy` | [codersdk.AIBridgeProxyConfig](#codersdkaibridgeproxyconfig) | false | | |
| `bridge` | [codersdk.AIBridgeConfig](#codersdkaibridgeconfig) | false | | |
| `chat` | [codersdk.ChatConfig](#codersdkchatconfig) | false | | |
## codersdk.APIAllowListTarget
@@ -1557,6 +1561,20 @@ AuthorizationObject can represent a "set" of objects, such as: all workspaces in
| `one_time_passcode` | string | true | | |
| `password` | string | true | | |
## codersdk.ChatConfig
```json
{
"acquire_batch_size": 0
}
```
### Properties
| Name | Type | Required | Restrictions | Description |
|----------------------|---------|----------|--------------|-------------|
| `acquire_batch_size` | integer | false | | |
## codersdk.ConnectionLatency
```json
@@ -2720,6 +2738,9 @@ CreateWorkspaceRequest provides options for creating a new workspace. Only one o
"retention": 0,
"send_actor_headers": true,
"structured_logging": true
},
"chat": {
"acquire_batch_size": 0
}
},
"allow_workspace_renames": true,
@@ -3292,6 +3313,9 @@ CreateWorkspaceRequest provides options for creating a new workspace. Only one o
"retention": 0,
"send_actor_headers": true,
"structured_logging": true
},
"chat": {
"acquire_batch_size": 0
}
},
"allow_workspace_renames": true,
site/src/api/typesGenerated.ts
Generated
+6
View File
@@ -136,6 +136,7 @@ export interface AIBridgeUserPrompt {
export interface AIConfig {
readonly bridge?: AIBridgeConfig;
readonly aibridge_proxy?: AIBridgeProxyConfig;
readonly chat?: ChatConfig;
}
// From codersdk/allowlist.go
@@ -1071,6 +1072,11 @@ export interface Chat {
readonly archived: boolean;
}
// From codersdk/deployment.go
export interface ChatConfig {
readonly acquire_batch_size: number;
}
// From codersdk/chats.go
/**
* ChatCostChatBreakdown contains per-root-chat cost aggregation.