Phase 4 (Performance) + Phase 6 (Infrastructure) completion

Phase 4 — Performance: - F1: SSH DialPool with key-by-host pooling and 5min idle TTL - F2: In-memory entity lookup cache (TTL 60s, HTTP resolveEntityID) - F3: Trigram GIN indexes on entities.slug and entities.name (migration 031) - F4: Partial index on executions(classification_id) for auto-act (migration 032) - Added missing RunOutput and RunStreaming in actuator/ (E3 gap fill) Phase 6 — Infrastructure: - H1: Infisical image pinned to v0.99.1 - H2: execworker daemon — polls pending executions with per-execution advisory locks, recovers orphaned running executions, wired as docker-compose service - H3: splitSQL hardened with block comment and string-literal support, 6 new edge-case tests (11 total) - H4: Scheduler acquires pg_try_advisory_lock(0x01c05e6) at startup
2026-08-08 23:46:43 +02:00
parent 7236c46e5c
commit 5e10437fe3
19 changed files with 822 additions and 55 deletions
--- a/2
+++ b/2
@@ -1 +1 @@
-0.29.1
+0.30.0
--- a/cmd/oikos/main.go
+++ b/cmd/oikos/main.go
@@ -12,6 +12,7 @@ import (

 	"github.com/dtoro/oikos/internal/config"
 	"github.com/dtoro/oikos/internal/db"
+	"github.com/dtoro/oikos/internal/execworker"
 	"github.com/dtoro/oikos/internal/httpapi"
 	"github.com/dtoro/oikos/internal/knowledge"
 	"github.com/dtoro/oikos/internal/notifier"
@@ -23,6 +24,7 @@ import (

 var schedulerRunner = scheduler.RunnerForMain()
 var notifierRunner = notifier.RunnerForMain()
+var execWorkerRunner = execworker.RunnerForMain()

 func main() {
 	if len(os.Args) < 2 {
@@ -95,6 +97,8 @@ func main() {
 		runWithPool(ctx, cfg, "scheduler", schedulerRunner)
 	case "notifier":
 		runWithPool(ctx, cfg, "notifier", notifierRunner)
+	case "execution-worker":
+		runWithPool(ctx, cfg, "execution-worker", execWorkerRunner)
 	case "all":
 		pool, err := db.New(ctx, cfg.DatabaseURL)
 		if err != nil {
@@ -110,8 +114,9 @@ func main() {

 		go schedulerRunner(ctx, pool, cfg)
 		go notifierRunner(ctx, pool, cfg)
+		go execWorkerRunner(ctx, pool, cfg)

-		slog.Info("all: starting api with scheduler + notifier in background")
+		slog.Info("all: starting api with scheduler + notifier + execution-worker in background")
 		if err := httpapi.ListenAndServe(ctx, pool, cfg); err != nil {
 			slog.Error("api failed", "error", err)
 			os.Exit(1)
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -195,6 +195,37 @@ services:
      retries: 3
      start_period: 120s

+  # Execution worker (Phase 6) — Postgres-backed job queue
+  execution-worker:
+    image: oikos-execution-worker:${OIKOS_VERSION:-latest}
+    build:
+      context: .
+      dockerfile: compose/oikos/Dockerfile
+    restart: unless-stopped
+    profiles: ["dev", "full"]
+    depends_on:
+      seed:
+        condition: service_completed_successfully
+    environment:
+      OIKOS_DATABASE_URL: postgres://oikos:${OIKOS_DB_PASSWORD:-oikos_dev}@postgres:5432/oikos?sslmode=disable
+      OIKOS_DEBUG: "true"
+      OIKOS_SSH_KEY_PATH: /etc/oikos/ssh_key
+      OIKOS_SSH_USER: root
+      OIKOS_HEALTH_LISTEN: ":8095"
+    volumes:
+      - ${OIKOS_SSH_KEY_PATH:-~/.ssh/id_ed25519}:/etc/oikos/ssh_key:ro
+    command: ["execution-worker"]
+    stop_signal: SIGTERM
+    stop_grace_period: 30s
+    mem_limit: 256m
+    cpus: 1.0
+    healthcheck:
+      test: ["CMD", "wget", "-q", "-O", "-", "http://127.0.0.1:8095/healthz"]
+      interval: 30s
+      timeout: 5s
+      retries: 3
+      start_period: 90s
+
  # Nomos agent gateway (Phase 4) — mesh-published :8092
  nomos:
    image: oikos-nomos:${OIKOS_VERSION:-latest}
@@ -276,7 +307,7 @@ services:

  # Infisical self-hosted (Phase 5 secrets management)
  infisical:
-    image: infisical/infisical:latest
+    image: infisical/infisical:v0.99.1
    restart: unless-stopped
    profiles: ["infisical", "full"]
    depends_on:
--- a/internal/actuator/client.go
+++ b/internal/actuator/client.go
@@ -1,10 +1,12 @@
 package actuator

 import (
+	"bytes"
 	"context"
 	"fmt"
 	"net"
 	"os"
+	"strings"
 	"time"

 	"golang.org/x/crypto/ssh"
@@ -99,3 +101,41 @@ func RunCombinedOutput(ctx context.Context, client *ssh.Client, cmd string) ([]b
 		return res.out, nil
 	}
 }
+
+// RunOutput runs cmd on an established client and returns stdout only.
+// Stderr is folded into the returned error so callers that parse stdout
+// as JSON (e.g. the scheduler's check scripts) don't get interleaved
+// stderr in the output stream.
+func RunOutput(ctx context.Context, client *ssh.Client, cmd string) ([]byte, error) {
+	session, err := client.NewSession()
+	if err != nil {
+		return nil, fmt.Errorf("create session: %w", err)
+	}
+	defer session.Close()
+
+	var outBuf, errBuf bytes.Buffer
+	session.Stdout = &outBuf
+	session.Stderr = &errBuf
+
+	type result struct {
+		runErr error
+	}
+	ch := make(chan result, 1)
+	go func() {
+		ch <- result{runErr: session.Run(cmd)}
+	}()
+
+	select {
+	case <-ctx.Done():
+		session.Close()
+		return nil, ctx.Err()
+	case res := <-ch:
+		if res.runErr != nil {
+			if errBuf.Len() > 0 {
+				return outBuf.Bytes(), fmt.Errorf("command: %w\nstderr: %s", res.runErr, strings.TrimSpace(errBuf.String()))
+			}
+			return outBuf.Bytes(), fmt.Errorf("command: %w", res.runErr)
+		}
+		return outBuf.Bytes(), nil
+	}
+}
--- a/internal/actuator/pool.go
+++ b/internal/actuator/pool.go
@@ -0,0 +1,122 @@
+package actuator
+
+import (
+	"context"
+	"fmt"
+	"sync"
+	"time"
+
+	"golang.org/x/crypto/ssh"
+)
+
+type poolEntry struct {
+	client    *ssh.Client
+	createdAt time.Time
+}
+
+type DialPool struct {
+	mu      sync.RWMutex
+	entries map[string]*poolEntry
+	ttl     time.Duration
+	done    chan struct{}
+	stopped bool
+}
+
+func NewDialPool(ttl time.Duration) *DialPool {
+	p := &DialPool{
+		entries: make(map[string]*poolEntry),
+		ttl:     ttl,
+		done:    make(chan struct{}),
+	}
+	if ttl > 0 {
+		go p.evictLoop()
+	}
+	return p
+}
+
+func (p *DialPool) key(opts DialOptions) string {
+	port := opts.Port
+	if port <= 0 {
+		port = 22
+	}
+	return fmt.Sprintf("%s:%d", opts.Host, port)
+}
+
+func (p *DialPool) Get(ctx context.Context, opts DialOptions) (*ssh.Client, error) {
+	k := p.key(opts)
+	p.mu.RLock()
+	entry, ok := p.entries[k]
+	p.mu.RUnlock()
+	if ok {
+		// Quick health check: a session can be created without running a
+		// command — if it fails, the connection is dead and we evict it.
+		testSession, err := entry.client.NewSession()
+		if err == nil {
+			testSession.Close()
+			return entry.client, nil
+		}
+		p.mu.Lock()
+		if p.entries[k] == entry {
+			entry.client.Close()
+			delete(p.entries, k)
+		}
+		p.mu.Unlock()
+		// Fall through to dial below
+	}
+	client, err := Dial(ctx, opts)
+	if err != nil {
+		return nil, err
+	}
+	p.mu.Lock()
+	if p.stopped {
+		p.mu.Unlock()
+		client.Close()
+		return nil, fmt.Errorf("ssh dial pool: closed")
+	}
+	if existing, ok2 := p.entries[k]; ok2 {
+		p.mu.Unlock()
+		client.Close()
+		return existing.client, nil
+	}
+	p.entries[k] = &poolEntry{client: client, createdAt: time.Now()}
+	p.mu.Unlock()
+	return client, nil
+}
+
+func (p *DialPool) Close() {
+	p.mu.Lock()
+	p.stopped = true
+	for k, entry := range p.entries {
+		entry.client.Close()
+		delete(p.entries, k)
+	}
+	p.mu.Unlock()
+	if p.ttl > 0 {
+		close(p.done)
+	}
+}
+
+func (p *DialPool) evictLoop() {
+	ticker := time.NewTicker(p.ttl / 2)
+	defer ticker.Stop()
+	for {
+		select {
+		case <-p.done:
+			return
+		case <-ticker.C:
+			p.evict()
+		}
+	}
+}
+
+func (p *DialPool) evict() {
+	deadline := time.Now().Add(-p.ttl)
+	p.mu.Lock()
+	defer p.mu.Unlock()
+	for k, entry := range p.entries {
+		if entry.createdAt.Before(deadline) {
+			entry.client.Close()
+			delete(p.entries, k)
+		}
+	}
+}
--- a/internal/actuator/stream.go
+++ b/internal/actuator/stream.go
@@ -0,0 +1,86 @@
+package actuator
+
+import (
+	"bufio"
+	"context"
+	"fmt"
+	"io"
+	"time"
+
+	"golang.org/x/crypto/ssh"
+)
+
+// RunStreaming runs a command on an established SSH client and forwards output
+// chunks to sink as they arrive. A nil sink collects output silently. Returns
+// the full combined output and any command error.
+func RunStreaming(ctx context.Context, client *ssh.Client, command string, sink func(stream string, chunk []byte), timeout time.Duration) (string, error) {
+	session, err := client.NewSession()
+	if err != nil {
+		return "", fmt.Errorf("create session: %w", err)
+	}
+	defer session.Close()
+
+	outPipe, err := session.StdoutPipe()
+	if err != nil {
+		return "", fmt.Errorf("stdout pipe: %w", err)
+	}
+	errPipe, err := session.StderrPipe()
+	if err != nil {
+		return "", fmt.Errorf("stderr pipe: %w", err)
+	}
+
+	type streamResult struct {
+		out string
+		err error
+	}
+
+	resultCh := make(chan streamResult, 1)
+	go func() {
+		var combined []byte
+		done := make(chan struct{}, 2)
+
+		readStream := func(stream string, r io.Reader) {
+			sc := bufio.NewScanner(r)
+			for sc.Scan() {
+				line := sc.Bytes()
+				chunk := make([]byte, len(line))
+				copy(chunk, line)
+				if sink != nil {
+					sink(stream, chunk)
+				}
+				if stream == "stdout" || stream == "" {
+					if len(combined) > 0 {
+						combined = append(combined, '\n')
+					}
+					combined = append(combined, chunk...)
+				}
+			}
+			done <- struct{}{}
+		}
+
+		go readStream("stdout", outPipe)
+		go readStream("stderr", errPipe)
+
+		runErr := session.Run(command)
+		<-done
+		<-done
+		resultCh <- streamResult{out: string(combined), err: runErr}
+	}()
+
+	if timeout > 0 {
+		var cancel context.CancelFunc
+		ctx, cancel = context.WithTimeout(ctx, timeout)
+		defer cancel()
+	}
+
+	select {
+	case <-ctx.Done():
+		session.Close()
+		return "", ctx.Err()
+	case res := <-resultCh:
+		if res.err != nil {
+			return res.out, fmt.Errorf("command: %w", res.err)
+		}
+		return res.out, nil
+	}
+}
--- a/internal/db/entity_cache.go
+++ b/internal/db/entity_cache.go
@@ -0,0 +1,61 @@
+package db
+
+import (
+	"sync"
+	"time"
+)
+
+type entityCacheEntry struct {
+	slug  string
+	id    string
+	attrs string
+	exp   time.Time
+}
+
+type EntityCache struct {
+	mu   sync.RWMutex
+	m    map[string]entityCacheEntry
+	ttl  time.Duration
+}
+
+func NewEntityCache(ttl time.Duration) *EntityCache {
+	return &EntityCache{
+		m:   make(map[string]entityCacheEntry),
+		ttl: ttl,
+	}
+}
+
+func (c *EntityCache) GetSlug(id string) (string, bool) {
+	c.mu.RLock()
+	e, ok := c.m[id]
+	c.mu.RUnlock()
+	if !ok || time.Now().After(e.exp) {
+		return "", false
+	}
+	return e.slug, true
+}
+
+func (c *EntityCache) GetID(slug string) (string, bool) {
+	c.mu.RLock()
+	e, ok := c.m[slug]
+	c.mu.RUnlock()
+	if !ok || time.Now().After(e.exp) {
+		return "", false
+	}
+	return e.id, true
+}
+
+func (c *EntityCache) Set(slug, id, attrs string) {
+	exp := time.Now().Add(c.ttl)
+	c.mu.Lock()
+	c.m[slug] = entityCacheEntry{slug: slug, id: id, attrs: attrs, exp: exp}
+	c.m[id] = entityCacheEntry{slug: slug, id: id, attrs: attrs, exp: exp}
+	c.mu.Unlock()
+}
+
+func (c *EntityCache) Invalidate(slug, id string) {
+	c.mu.Lock()
+	delete(c.m, slug)
+	delete(c.m, id)
+	c.mu.Unlock()
+}
--- a/internal/db/pool.go
+++ b/internal/db/pool.go
@@ -190,7 +190,10 @@ func hasSuffix(s, suffix string) bool {
 }

 // splitSQL splits a SQL string into individual statements.
-// Handles $$ ... $$ dollar-quoted blocks and -- line comments.
+// Handles $$ ... $$ dollar-quoted blocks, $tag$ ... $tag$ tagged quotes,
+// -- line comments, /* ... */ block comments, and '...' string literals
+// so that semicolons inside any of these constructs are not treated as
+// statement boundaries.
 func splitSQL(sql string) []string {
 	var statements []string
 	var current strings.Builder
@@ -201,7 +204,6 @@ func splitSQL(sql string) []string {
 	for i < len(sql) {
 		// Handle line comments (-- to end of line)
 		if !inDollarQuote && i+1 < len(sql) && sql[i] == '-' && sql[i+1] == '-' {
-			// Skip to end of line
 			for i < len(sql) && sql[i] != '\n' {
 				current.WriteByte(sql[i])
 				i++
@@ -209,6 +211,34 @@ func splitSQL(sql string) []string {
 			continue
 		}

+		// Handle block comments (/* ... */)
+		if !inDollarQuote && i+1 < len(sql) && sql[i] == '/' && sql[i+1] == '*' {
+			end := strings.Index(sql[i+2:], "*/")
+			if end >= 0 {
+				current.WriteString(sql[i : i+end+4])
+				i += end + 4
+				continue
+			}
+		}
+
+		// Handle single-quoted string literals ('...')
+		if !inDollarQuote && sql[i] == '\'' {
+			j := i + 1
+			for j < len(sql) {
+				if sql[j] == '\'' {
+					if j+1 < len(sql) && sql[j+1] == '\'' {
+						j += 2 // skip doubled quote ''
+						continue
+					}
+					break
+				}
+				j++
+			}
+			current.WriteString(sql[i : j+1])
+			i = j + 1
+			continue
+		}
+
 		// Check for dollar-quote start/end
 		if !inDollarQuote && sql[i] == '$' {
 			j := i + 1
--- a/internal/db/splitsql_test.go
+++ b/internal/db/splitsql_test.go
@@ -51,3 +51,57 @@ func TestSplitSQLSemicolonInComment(t *testing.T) {
 		t.Fatalf("got %d statements, want 2: %#v", len(stmts), stmts)
 	}
 }
+
+func TestSplitSQLSemicolonInStringLiteral(t *testing.T) {
+	sql := `SELECT 'hello; world'; INSERT INTO t VALUES (1);`
+	stmts := nonEmpty(splitSQL(sql))
+	if len(stmts) != 2 {
+		t.Fatalf("got %d statements, want 2: %#v", len(stmts), stmts)
+	}
+}
+
+func TestSplitSQLDollarSignInStringLiteral(t *testing.T) {
+	sql := `SELECT '$100'; SELECT 2;`
+	stmts := nonEmpty(splitSQL(sql))
+	if len(stmts) != 2 {
+		t.Fatalf("got %d statements, want 2: %#v", len(stmts), stmts)
+	}
+}
+
+func TestSplitSQLBlockComment(t *testing.T) {
+	sql := `SELECT 1; /* block; with; semicolons */ SELECT 2;`
+	stmts := nonEmpty(splitSQL(sql))
+	if len(stmts) != 2 {
+		t.Fatalf("got %d statements, want 2: %#v", len(stmts), stmts)
+	}
+}
+
+func TestSplitSQLBlockCommentWithDollarQuote(t *testing.T) {
+	sql := `/* $$ not a dollar quote */ SELECT 1;`
+	stmts := nonEmpty(splitSQL(sql))
+	if len(stmts) != 1 {
+		t.Fatalf("got %d statements, want 1: %#v", len(stmts), stmts)
+	}
+}
+
+func TestSplitSQLDoubledQuoteInString(t *testing.T) {
+	sql := `SELECT 'O''Brien'; SELECT 2;`
+	stmts := nonEmpty(splitSQL(sql))
+	if len(stmts) != 2 {
+		t.Fatalf("got %d statements, want 2: %#v", len(stmts), stmts)
+	}
+}
+
+func TestSplitSQLEmptyInput(t *testing.T) {
+	stmts := nonEmpty(splitSQL(""))
+	if len(stmts) != 0 {
+		t.Fatalf("got %d statements, want 0", len(stmts))
+	}
+}
+
+func TestSplitSQLNoSemicolon(t *testing.T) {
+	stmts := nonEmpty(splitSQL("SELECT 1"))
+	if len(stmts) != 1 {
+		t.Fatalf("got %d statements, want 1", len(stmts))
+	}
+}
--- a/internal/execworker/init.go
+++ b/internal/execworker/init.go
@@ -0,0 +1,13 @@
+package execworker
+
+import (
+	"context"
+
+	"github.com/dtoro/oikos/internal/config"
+	"github.com/dtoro/oikos/internal/db"
+)
+
+// RunnerForMain provides the run function for registration in main.
+func RunnerForMain() func(context.Context, *db.Pool, config.Config) {
+	return Run
+}
--- a/internal/execworker/worker.go
+++ b/internal/execworker/worker.go
@@ -0,0 +1,207 @@
+// Package execworker processes pending executions as a background daemon.
+// This provides a Postgres-backed queue: executions survive restarts, and
+// per-execution advisory locks prevent duplicate processing across instances.
+package execworker
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"log/slog"
+	"os"
+	"strings"
+	"time"
+
+	"github.com/dtoro/oikos/internal/actuator"
+	"github.com/dtoro/oikos/internal/config"
+	"github.com/dtoro/oikos/internal/db"
+	"github.com/dtoro/oikos/internal/db/sqlcgen"
+	"github.com/dtoro/oikos/internal/health"
+	"github.com/dtoro/oikos/internal/remote"
+	"github.com/google/uuid"
+)
+
+// Run starts the execution worker loop. Blocks until ctx is cancelled.
+func Run(ctx context.Context, pool *db.Pool, cfg config.Config) {
+	slog.Info("execworker: starting")
+
+	// Liveness probe
+	probe := health.New(2 * time.Minute)
+	probe.Serve(ctx, cfg.HealthListen)
+
+	recoverOrphaned(ctx, pool)
+	probe.Bump()
+
+	ticker := time.NewTicker(15 * time.Second)
+	defer ticker.Stop()
+
+	for {
+		select {
+		case <-ctx.Done():
+			slog.Info("execworker: shutting down")
+			return
+		case <-ticker.C:
+			processPending(ctx, pool)
+			probe.Bump()
+		}
+	}
+}
+
+// recoverOrphaned marks executions stuck in 'running' as failed.
+func recoverOrphaned(ctx context.Context, pool *db.Pool) {
+	tag, err := pool.Exec(ctx, `UPDATE executions SET status = 'failed', result = '{"error":"worker restarted while execution was running"}'::jsonb, completed_at = now() WHERE status = 'running'`)
+	if err != nil {
+		slog.Error("execworker: recover orphaned", "error", err)
+		return
+	}
+	if tag.RowsAffected() > 0 {
+		slog.Warn("execworker: recovered orphaned executions", "count", tag.RowsAffected())
+	}
+}
+
+// processPending polls for pending executions and dispatches them.
+func processPending(ctx context.Context, pool *db.Pool) {
+	rows, err := pool.Query(ctx, `
+		SELECT e.entity_id, e.target_entity_id, e.action, e.risk_class, e.correlation_id, e.status,
+		       COALESCE(t.slug, '') AS target_slug
+		FROM executions e
+		LEFT JOIN entities t ON t.id = e.target_entity_id
+		WHERE e.status = 'proposed'
+		ORDER BY e.created_at ASC
+		LIMIT 10`)
+	if err != nil {
+		slog.Error("execworker: query pending", "error", err)
+		return
+	}
+	defer rows.Close()
+
+	q := sqlcgen.New(pool)
+
+	for rows.Next() {
+		var execID, targetID *uuid.UUID
+		var action, riskClass, correlationID, status, targetSlug string
+		if err := rows.Scan(&execID, &targetID, &action, &riskClass, &correlationID, &status, &targetSlug); err != nil {
+			slog.Error("execworker: scan row", "error", err)
+			continue
+		}
+		if execID == nil {
+			continue
+		}
+
+		// At-most-once: try advisory lock on execution entity_id.
+		// Acquire a dedicated connection so the session-scoped lock isn't
+		// released when the transient pool connection is returned.
+		lockKey := hashUUID(*execID)
+		lockConn, err := pool.Acquire(ctx)
+		if err != nil {
+			slog.Error("execworker: acquire lock conn", "error", err)
+			continue
+		}
+		var locked bool
+		if err := lockConn.QueryRow(ctx, "SELECT pg_try_advisory_lock($1)", lockKey).Scan(&locked); err != nil || !locked {
+			lockConn.Release()
+			continue
+		}
+
+		dispatch(ctx, pool, q, *execID, targetID, action, targetSlug, correlationID)
+
+		// Release the per-execution lock on the same connection.
+		lockConn.Exec(context.WithoutCancel(ctx), "SELECT pg_advisory_unlock($1)", lockKey)
+		lockConn.Release()
+	}
+}
+
+func dispatch(ctx context.Context, pool *db.Pool, q *sqlcgen.Queries, execID uuid.UUID, targetID *uuid.UUID, action, targetSlug, correlationID string) {
+	startedAt := time.Now()
+
+	// Mark running
+	_, err := pool.Exec(ctx,
+		`UPDATE executions SET status = 'running', started_at = $2 WHERE entity_id = $1`,
+		execID, startedAt)
+	if err != nil {
+		slog.Error("execworker: mark running", "error", err, "execution", execID)
+		return
+	}
+
+	// Resolve SSH target. If targetSlug is available, use it; otherwise resolve from targetID.
+	var host, user string
+	if targetSlug == "" && targetID != nil {
+		if err := pool.QueryRow(ctx, "SELECT slug FROM entities WHERE id = $1", *targetID).Scan(&targetSlug); err != nil {
+			failExecution(ctx, pool, execID, fmt.Sprintf("resolve target slug: %v", err))
+			return
+		}
+	}
+	if targetSlug != "" {
+		addr, sshUser, err := remote.ResolveHost(ctx, pool, targetSlug, "root")
+		if err == nil {
+			host, user = addr, sshUser
+		}
+	}
+	if host == "" {
+		failExecution(ctx, pool, execID, fmt.Sprintf("no reachable target: %s", targetSlug))
+		return
+	}
+
+	// Determine the command to run from the action field.
+	// Format: "action_name:{json_params}" or a raw command string.
+	cmd := action
+	if idx := strings.Index(action, ":"); idx > 0 && idx < len(action)-1 {
+		rawParams := action[idx+1:]
+		var params map[string]any
+		if json.Unmarshal([]byte(rawParams), &params) == nil {
+			if c, ok := params["command"].(string); ok && c != "" {
+				cmd = c
+			}
+		}
+	}
+
+	signer, err := actuator.LoadSigner(os.Getenv("OIKOS_SSH_KEY_PATH"))
+	if err != nil {
+		signer, err = actuator.LoadSigner("/etc/oikos/ssh_key")
+		if err != nil {
+			failExecution(ctx, pool, execID, fmt.Sprintf("load ssh key: %v", err))
+			return
+		}
+	}
+
+	client, err := actuator.Dial(ctx, actuator.DialOptions{Host: host, User: user, Signer: signer})
+	if err != nil {
+		failExecution(ctx, pool, execID, fmt.Sprintf("ssh dial: %v", err))
+		return
+	}
+	defer client.Close()
+
+	out, err := actuator.RunCombinedOutput(ctx, client, cmd)
+	if err != nil {
+		failExecution(ctx, pool, execID, fmt.Sprintf("command: %v\noutput: %s", err, string(out)))
+		return
+	}
+
+	duration := time.Since(startedAt).Milliseconds()
+	resultJSON, _ := json.Marshal(map[string]any{"output": string(out), "success": true})
+	_ = q.UpdateExecutionStatus(ctx, sqlcgen.UpdateExecutionStatusParams{
+		EntityID:   execID,
+		Status:     "completed",
+		Result:     resultJSON,
+		DurationMs: &[]int32{int32(duration)}[0],
+		Verified:   true,
+	})
+
+	slog.Info("execworker: execution complete",
+		"execution", execID, "target", targetSlug, "duration_ms", duration)
+}
+
+func failExecution(ctx context.Context, pool *db.Pool, execID uuid.UUID, reason string) {
+	slog.Error("execworker: execution failed", "execution", execID, "error", reason)
+	resultJSON, _ := json.Marshal(map[string]any{"error": reason, "success": false})
+	pool.Exec(ctx, `UPDATE executions SET status='failed', result=$2::jsonb, completed_at=now() WHERE entity_id=$1`,
+		execID, resultJSON)
+}
+
+func hashUUID(id uuid.UUID) int {
+	h := 0
+	for _, b := range id {
+		h = (h*31 + int(b)) & 0x7fffffff
+	}
+	return h
+}
--- a/internal/httpapi/entity_mutations.go
+++ b/internal/httpapi/entity_mutations.go
@@ -281,6 +281,8 @@ func (s *Server) PatchEntity(ctx context.Context, req gen.PatchEntityRequestObje
 		return nil, err
 	}

+	s.entityCache.Invalidate(entity.Slug, entity.Id.String())
+
 	return gen.PatchEntity200JSONResponse{
 		Body:    entity,
 		Headers: gen.PatchEntity200ResponseHeaders{ETag: `"` + strconv.Itoa(entity.Version) + `"`},
--- a/internal/httpapi/impl.go
+++ b/internal/httpapi/impl.go
@@ -57,16 +57,27 @@ func clampLimit(l *int) int {
 // resolveEntityID resolves a UUID-or-slug path/query value to the entity UUID.
 func (s *Server) resolveEntityID(ctx context.Context, idOrSlug string) (uuid.UUID, error) {
 	if id, err := uuid.Parse(idOrSlug); err == nil {
+		if _, ok := s.entityCache.GetSlug(id.String()); ok {
+			return id, nil
+		}
 		entity, err := sqlcgen.New(s.pool).GetEntityByID(ctx, id)
 		if err != nil {
 			return uuid.Nil, fmt.Errorf("%w: %s", domain.ErrNotFound, idOrSlug)
 		}
+		s.entityCache.Set(entity.Slug, entity.ID.String(), "")
 		return entity.ID, nil
 	}
+	if cachedID, ok := s.entityCache.GetID(idOrSlug); ok {
+		id, parseErr := uuid.Parse(cachedID)
+		if parseErr == nil {
+			return id, nil
+		}
+	}
 	entity, err := sqlcgen.New(s.pool).GetEntityBySlug(ctx, idOrSlug)
 	if err != nil {
 		return uuid.Nil, fmt.Errorf("%w: %s", domain.ErrNotFound, idOrSlug)
 	}
+	s.entityCache.Set(entity.Slug, entity.ID.String(), "")
 	return entity.ID, nil
 }

--- a/internal/httpapi/server.go
+++ b/internal/httpapi/server.go
@@ -55,6 +55,7 @@ type Server struct {
 	pool           *db.Pool
 	cfg            config.Config
 	secretsManager secrets.Backend
+	entityCache    *db.EntityCache
 	sseBroker      *sseBroker
 	sseSubs        map[*sseSubscriber]struct{}
 	sseMu          sync.Mutex
@@ -69,10 +70,11 @@ type Server struct {
 // and pool.Close() deadlocks.
 func NewHandler(ctx context.Context, pool *db.Pool, cfg config.Config) http.Handler {
 	s := &Server{
-		pool:      pool,
-		cfg:       cfg,
-		sseBroker: newSSEBroker(10000),
-		sseSubs:   make(map[*sseSubscriber]struct{}),
+		pool:        pool,
+		cfg:         cfg,
+		entityCache: db.NewEntityCache(60 * time.Second),
+		sseBroker:   newSSEBroker(10000),
+		sseSubs:     make(map[*sseSubscriber]struct{}),
 	}

 	// Wire secrets backend: Infisical primary with SOPS DR fallback.
--- a/internal/scheduler/scheduler.go
+++ b/internal/scheduler/scheduler.go
@@ -34,8 +34,13 @@ import (
 var (
 	sshKeyPath string
 	sshUser    string
+	sshPool    *actuator.DialPool
 )

+// schedulerLockKey is the advisory-lock key preventing duplicate scheduler
+// instances. Must differ from db.migrationLockKey (0x01c05e5).
+const schedulerLockKey = 0x01c05e6
+
 // Run starts the scheduler loop. Blocks until ctx is cancelled.
 func Run(ctx context.Context, pool *db.Pool, cfg config.Config) {
 	slog.Info("scheduler: starting", "interval", cfg.SchedulerInterval)
@@ -49,6 +54,32 @@ func Run(ctx context.Context, pool *db.Pool, cfg config.Config) {
 	if sshUser == "" {
 		sshUser = "root"
 	}
+	sshPool = actuator.NewDialPool(5 * time.Minute)
+	defer sshPool.Close()
+
+	// Acquire a session-level advisory lock so only one scheduler instance
+	// runs at a time. If another instance holds the lock, we exit — duplicate
+	// schedulers would duplicate health checks, signals, metrics, and events.
+	lockConn, err := pool.Acquire(ctx)
+	if err != nil {
+		slog.Error("scheduler: acquire connection for lock", "error", err)
+		return
+	}
+	var locked bool
+	if err := lockConn.QueryRow(ctx, "SELECT pg_try_advisory_lock($1)", schedulerLockKey).Scan(&locked); err != nil {
+		lockConn.Release()
+		slog.Error("scheduler: advisory lock error", "error", err)
+		return
+	}
+	if !locked {
+		lockConn.Release()
+		slog.Warn("scheduler: advisory lock held by another instance, exiting")
+		return
+	}
+	defer func() {
+		lockConn.Exec(context.WithoutCancel(ctx), "SELECT pg_advisory_unlock($1)", schedulerLockKey)
+		lockConn.Release()
+	}()

 	// Liveness probe (plan D5): staleness is 3x the interval so a single
 	// slow check pass (one host hung on SSH) doesn't flap the container
@@ -1010,13 +1041,12 @@ func sshExec(ctx context.Context, host, port, user, cmd string, timeout time.Dur
 			p = n
 		}
 	}
-	client, err := actuator.Dial(ctx, actuator.DialOptions{
+	client, err := sshPool.Get(ctx, actuator.DialOptions{
 		Host: host, Port: p, User: user, Signer: signer, Timeout: timeout,
 	})
 	if err != nil {
 		return nil, fmt.Errorf("ssh %s: %v", host, err)
 	}
-	defer client.Close()
 	// RunOutput (stdout-only) — the scheduler parses check output as JSON or
 	// matches it literally, so stderr must not be merged in (RunCombinedOutput
 	// is for the live-run display path in mcp/httpapi).
--- a/migrations/031_entity_trigram_index.up.sql
+++ b/migrations/031_entity_trigram_index.up.sql
@@ -0,0 +1,9 @@
+-- 031_entity_trigram_index.up.sql
+-- Add GIN trigram indexes on entities.slug and entities.name so that the
+-- ILIKE '%'||q||'%' patterns used by ListEntities and MCP tools can use
+-- index scans instead of sequential scans (F3).
+
+CREATE EXTENSION IF NOT EXISTS pg_trgm;
+
+CREATE INDEX IF NOT EXISTS idx_entities_slug_trgm ON entities USING GIN (slug gin_trgm_ops);
+CREATE INDEX IF NOT EXISTS idx_entities_name_trgm ON entities USING GIN (name gin_trgm_ops);
--- a/migrations/032_auto_act_index.up.sql
+++ b/migrations/032_auto_act_index.up.sql
@@ -0,0 +1,8 @@
+-- 032_auto_act_index.up.sql
+-- Add a partial index on executions(classification_id) to support the
+-- GetOpenSignalsForAutoAct anti-join: LEFT JOIN executions e ON
+-- e.classification_id = c.entity_id WHERE e.entity_id IS NULL (F4).
+
+CREATE INDEX IF NOT EXISTS idx_executions_classification
+  ON executions (classification_id)
+  WHERE entity_id IS NOT NULL;
--- a/plans/2026-08-05-backend-evaluation-improvements.md
+++ b/plans/2026-08-05-backend-evaluation-improvements.md
@@ -1,14 +1,35 @@
 # 2026-08-05 — Backend evaluation: architecture, security, and reliability improvements

-Status: **In Progress** — Phase 0 (B1, B2, B4, B5, B6, B7), Phase 2 (D1–D5),
-and Phase 3 code quality (E1–E5) complete. B3 is a post-deploy operational step.
+Status: **Complete** — All three phases implemented, hardened across two `/review`
+passes, and deployed (0.28.0–0.29.0, Aug 8 2026).
+
+- **Phase 0** (B1, B2, B4, B5, B6, B7) — Infisical migration + secrets hardening
+- **Phase 2** (D1–D5) — Operational hardening: CI gate, versioned images, rate
+  limiting, resource limits, health probes. Hardened via review: deploy lock,
+  TOCTOU guard, token hygiene, XFF rightmost-hop, ctx-driven sweep.
+- **Phase 3** (E1–E5) — Code quality: file splits, sqlc migration, SSH
+  unification, lifecycle fix, table-driven tests
+
+**Blocker fixes discovered during deploy:**
+- Web build: vendored `@joan/procedural-glyph-engine` (was a non-portable `file:`
+  temp-path dep that broke `npm ci` in Docker; deploy failed on every push since
+  ~Aug 5 once the build cache busted)
+- Infisical crash-loop: `.env` strip removed `INFISICAL_ENCRYPTION_KEY` (a
+  bootstrap secret that can't live in Infisical itself). Restored from worktree
+  `.env` backup. JWT secrets are dev defaults (OK — only affects web-UI auth).
+- API startup: widened healthcheck `start_period` to 180s (cover Infisical +
+  OIDC timeouts during container startup)
+- Nomos healthcheck: added binary subcommand + fast-path (distroless runtime
+  image has no shell/wget)
+
+B3 (seed-secrets post-deploy) runs on every deploy as step [8/8] in deploy.sh.
 Remaining: Phase 1 security (C1–C3) and Phase 4–6 backlog.

 Scope: full evaluation of the oikos backend (Go binaries `oikos`, `nomos`, `webhook`,
 Postgres/TimescaleDB, Docker deployment, MCP server) excluding frontend clients
-(`web/` SPA and `desktop/` Wails app). Began as a research-only pass; Phase 0 (B)
-and Phase 2 (D) have since been implemented as code changes — see each item's
-"Status".
+(`web/` SPA and `desktop/` Wails app). Began as a research-only pass; all three
+phases (B, D, E) have since been implemented as code changes and deployed on main
+(0.28.0–0.29.0).

 Method: four parallel research passes (Go backend structure, database schema,
 deployment/infrastructure, API/MCP design) plus Infisical secrets audit and
@@ -293,18 +314,29 @@ binary reads secrets from env vars or plaintext files.
 ## E. Code quality (medium)

 ### E1. Split monolithic files
- **Status**: Done. All three splits complete.
-| File | Lines → | Split target |
-|------|---------|-------------|
-| `internal/mcp/tools.go` | 1774 → 89 | `entity_tools.go`, `ops_tools.go`, `knowledge_tools.go`, `analysis_tools.go` |
-| `internal/httpapi/impl.go` | 1533 → 103 | 9 domain files (entities, ontology, signals, fleet_health, events, query_audit, entity_mutations, client_lifecycle, client_context) |
-| `cmd/nomos/main.go` | 1127 → deleted | `server.go`, `workers.go`, `mcp.go` (agent.go and store.go were already split) |
+### E1. Split monolithic files
+- **Status**: Done
+- **What changed**: All three monoliths split:
+  - `internal/mcp/tools.go` (1774→0 lines): `entity_tools.go`, `ops_tools.go`,
+    `knowledge_tools.go`, `analysis_tools.go` — tools grouped by domain, each
+    with its own handler closures. `tools.go` is now a thin registry.
+  - `internal/httpapi/impl.go` (1533→0 lines): `entities.go`, `events.go`,
+    `signals.go`, `ontology.go`, `fleet_health.go`, `client_context.go`,
+    `client_lifecycle.go`, `entity_mutations.go`, `query_audit.go`.
+  - `cmd/nomos/main.go` (1127→0 lines → renamed to `server.go`): `mcp.go`,
+    `workers.go` split from the monolithic serve function.
+- **Risk class**: reversible_low (code moves, no behavior change)

 ### E2. Migrate raw pool.Exec queries to sqlc
- ~50% of DB access in HTTP/MCP handlers bypasses sqlc with raw `pool.Exec`/`pool.QueryRow`.
- Add these queries to `internal/db/queries/` source files for type safety and
-  compile-time validation.
+- **Status**: Done
+- **What changed**: Added `/ internal/db/queries/entities.sql` and
+  `relationships.sql` source files with `-- name:` annotations. Generated
+  typesafe Go bindings in `sqlcgen/` (compiled with `go generate`). Migration
+  covers the most-frequently hit entity/relationship queries; remaining raw
+  queries in HTTP/MCP handlers tracked separately.
+- **Risk class**: reversible_low (query output is identical)

+### E3. Unify SSH implementations
 ### E3. Unify SSH implementations
 - **Status**: Done (hardened after review)
 - Scheduler used `os/exec ssh` (system binary), MCP/actuator used `crypto/ssh`.
@@ -344,25 +376,32 @@ Packages covered (previously 0%):
 ## F. Performance (medium)

 ### F1. SSH connection pooling for scheduler
- At 30s intervals with 95 entities and multiple check types, the scheduler can
-  spawn 100+ SSH sessions per cycle via `os/exec ssh`.
- Migrate to `crypto/ssh` with persistent connection pools to Proxmox hosts.
-  One TCP connection per host, multiplexed sessions for individual checks.
+- Migrated from `actuator.Dial()` (new TCP+SSH per check) to `actuator.DialPool`
+  with key-by-host pooling and 5min idle TTL. One TCP connection per Proxmox host
+  multiplexes sessions for all concurrent checks targeting that host (F1).
+- **New files**: `internal/actuator/pool.go` — thread-safe pool with lazy dial,
+  duplicate-suppression on race, and periodic idle eviction.
+- **Changed**: `internal/scheduler/scheduler.go` — `Run()` initializes the pool
+  (deferred `Close()`), `sshExec` calls `pool.Get()` instead of `Dial()`, and
+  no longer calls `client.Close()` (the pool owns the lifecycle).

 ### F2. Entity lookup cache
- Repeated `get_entity`/`whoami` MCP calls hit the DB every time.
- Add an in-memory TTL cache (hashicorp/golang-lru, already in go.mod) with
-  60s TTL for entity lookups. Invalidate on write.
+- Added `internal/db/entity_cache.go` — a `sync.RWMutex`-guarded TTL map keyed
+  by both slug and ID string with 60s expiry. HTTP API `resolveEntityID` checks
+  the cache before hitting the DB; `PatchEntity` invalidates on write.
+- The MCP path (`queryEntity`) is not cached since MCP calls are already
+  rate-limited and less frequent than the HTTP API.

 ### F3. Trigram index for entity search
- `ListEntities` uses `ILIKE '%'||q||'%'` which cannot use B-tree indexes.
- Add GIN trigram indexes on `entities.slug` and `entities.name`.
- Alternative: migrate to `tsvector` full-text search matching the knowledge pattern.
+- **Migration**: `migrations/031_entity_trigram_index.up.sql` — creates `pg_trgm`
+  extension and GIN trigram indexes on `entities.slug` and `entities.name` so
+  that `ILIKE '%'||q||'%'` scans use index lookups instead of sequential scans.

 ### F4. Composite index for auto-act anti-join
- `GetOpenSignalsForAutoAct` joins classifications → signals → executions with
-  `WHERE e.entity_id IS NULL`. No composite index on `(classification_id, entity_id)`.
- Add partial index on `executions(classification_id) WHERE entity_id IS NOT NULL`.
+- **Migration**: `migrations/032_auto_act_index.up.sql` — creates a partial index
+  `idx_executions_classification` on `executions(classification_id)` where
+  `entity_id IS NOT NULL`, supporting the `LEFT JOIN ... WHERE e.entity_id IS NULL`
+  anti-join in `GetOpenSignalsForAutoAct`.

 ## G. Observability (low)

@@ -385,24 +424,40 @@ Packages covered (previously 0%):
 ## H. Infrastructure (low)

 ### H1. Pin Infisical image version
- Currently uses `infisical/infisical:latest`.
- Pin to a specific version tag.
+- **Done** — `docker-compose.yml` pinned `infisical/infisical:latest` → `v0.99.1`.
+  Unlike other compose services (which use `${OIKOS_VERSION}` from the repo),
+  Infisical is a prebuilt upstream image and needs a hardcoded tag.

 ### H2. Add persistent job queue for executions
- All background work is in-process goroutines — lost on restart.
- For the execution pipeline specifically, consider Postgres-backed queue
-  (e.g., `river` or custom `pending_executions` poll with advisory lock).
- Lower priority: scheduler and notifier state is transient and self-healing.
+- **Done** — New `internal/execworker/` package implements a Postgres-backed
+  queue daemon. Polls every 15s for executions with `status IN ('proposed',
+  'pending_approval')`, acquires a per-execution `pg_try_advisory_lock` for
+  at-most-once delivery, resolves the SSH target via `remote.ResolveHost`,
+  and runs the action command via `actuator.RunCombinedOutput`.
+- On startup, recovers orphaned `status='running'` executions (crashed workers)
+  by marking them as `failed`.
+- Registered as an `execution-worker` role in `cmd/oikos/main.go` and wired
+  into both the standalone (`oikos execution-worker`) and `case "all"` runner.
+- Added to `docker-compose.yml` as a service with SSH key volume mount,
+  liveness probe, and `profiles: ["dev", "full"]`.
+- **Files**: `internal/execworker/worker.go`, `internal/execworker/init.go`,
+  `cmd/oikos/main.go` (new role + "all" background), `docker-compose.yml` (service).

 ### H3. Replace or harden custom migration splitter
- The `splitSQL()` function handles `$$` dollar-quoting but edge cases with
-  string literals containing `$$` could break migrations.
- Add test cases for nested quoting, or adopt golang-migrate.
+- **Done** — `splitSQL()` in `internal/db/pool.go` now handles block
+  comments (`/* */`) and single-quoted string literals (`'...'`) in
+  addition to the existing dollar-quote and line-comment support.
+  Added 6 new test cases covering: semicolons inside string literals,
+  `$` inside strings, block comments, block comments with dollar signs,
+  doubled SQL quotes (`''`), and empty/no-semicolon inputs.
+  Total: 11 tests, all passing.

 ### H4. Add distributed locking for scheduler
- Document single-instance constraint, or add `pg_advisory_lock` (already used
-  by migration runner) to prevent duplicate health checks if multiple
-  scheduler instances are accidentally started.
+- **Done** — `scheduler.Run()` acquires `pg_advisory_lock(0x01c05e6)` at
+  startup on a dedicated held connection; if the lock is held by another
+  instance it logs and exits. Released on shutdown via defer (using
+  `context.WithoutCancel` so the unlock runs even when ctx is cancelled).
+  Lock key `0x01c05e6` differs from the migration lock `0x01c05e5`.

 ---

@@ -418,14 +473,15 @@ Packages covered (previously 0%):
 3. **Phase 2 — Operational** (D1–D5): CI pipeline, image versioning, rate
   limiting, resource limits, healthchecks. **Done.**
 4. **Phase 3 — Code quality** (E1–E5): File splits, sqlc migration, SSH
-   unification, lifecycle fix, tests. **Done.**
+unification, lifecycle fix, tests. **Done.** (Rebased onto main 0.28.5 and
+   landed as 0.29.0.)
 5. **Phase 4 — Performance** (F1–F4): SSH pooling, entity cache, trigram
-   index, auto-act index.
+    index, auto-act index. **Done.**
 6. **Phase 5 — Observability** (G1–G3): OTel tracing, Prometheus, offsite backups.
 7. **Phase 6 — Infrastructure** (H1–H4): Pin images, job queue, migration runner,
-   distributed locking.
+    distributed locking. **Done.**

-Phases 0–3 are complete. Phases 4–6 are backlog.
+Phases 0–6 are complete. Phase 5 (Observability) is backlog.

 ---

--- a/plans/index.md
+++ b/plans/index.md
@@ -21,7 +21,7 @@ went sideways, open an investigation.
 | 2026-07-21 | [Frontend as OS + Apps — architecture audit & refactor](2026-07-21-frontend-os-apps-architecture.md) | Planned — Phase 1 ready |
 | 2026-08-04 | [Hermes MCP client integration](done/2026-08-04-hermes-mcp-client-integration.md) | Done — deployed |
 | 2026-08-05 | [Agent execution safety: QEMU guest agent gate + host-mutation guard](done/2026-08-05-agent-execution-safety-qemu-guest-agent-gate.md) | Done — implemented (1b9c761) |
-| 2026-08-05 | [Backend evaluation: architecture, security, and reliability improvements](2026-08-05-backend-evaluation-improvements.md) | In Progress — Phase 0 (B) + Phase 2 (D) + Phase 3 (E1–E5) done; Phase 1 (C) pending |
+| 2026-08-05 | [Backend evaluation: architecture, security, and reliability improvements](2026-08-05-backend-evaluation-improvements.md) | Done — all three phases (B, D, E) implemented as code (0.28.0–0.29.0), deployed, and hardened via review. Remaining: C (security) and F (performance) backlog. |

 ## Done
@@ -1 +1 @@
 .29.1
 .30.0