t8y2/dbx · error

agent request timed out

Error message

agent request timed out

What it means

An agent RPC that does not receive a response within a fixed 30-second window times out; the pending request is removed and this error returned. It indicates the agent process is hung, dead, or too slow rather than returning an error reply.

Source

Thrown at agents/drivers/kingbase-go/bench/agent_compare.go:249

	if err != nil {
		return nil, err
	}
	process.writeMu.Lock()
	_, err = process.stdin.Write(append(payload, '\n'))
	process.writeMu.Unlock()
	if err != nil {
		process.pending.Delete(id)
		return nil, err
	}
	select {
	case response := <-channel:
		if response.Error != nil {
			return nil, errors.New(response.Error.Message)
		}
		return response.Result, nil
	case <-time.After(30 * time.Second):
		process.pending.Delete(id)
		return nil, errors.New("agent request timed out")
	}
}

func (process *agentProcess) close() error {
	_, _ = process.call("shutdown", map[string]any{})
	_ = process.stdin.Close()
	return process.command.Wait()
}

func runQueryBenchmark(process *agentProcess, duration time.Duration, concurrency int, sqlText string, maxRows int, sampleMemory bool) benchmarkResult {
	var operations atomic.Int64
	var failures atomic.Int64
	var peakRSS atomic.Int64
	stopMemory := make(chan struct{})
	if sampleMemory {
		peakRSS.Store(readRSSKB(process.command.Process.Pid))
		go func() {
			ticker := time.NewTicker(100 * time.Millisecond)

View on GitHub (pinned to c0390bff16)

Solutions

1. Check whether the agent process is alive (logs, exit status) and restart it if it crashed or hung
2. Reduce query workload or increase the 30s timeout constant for legitimately long operations
3. Ensure the agent always writes exactly one response per request, even on internal errors

Example fix

// before
case <-time.After(30 * time.Second):
// after
timeout := 60 * time.Second // configurable for long-running queries
case <-time.After(timeout):

if process.exited() { return errors.New("agent process is not running; restart before calling") }

res, err := process.call(method, args); if err != nil && err.Error() == "agent request timed out" { // restart agent and retry once with a longer deadline
}

When it happens

Trigger: Calling process.call (from close, warmup, benchmarkOneConnection, cleanupQueryPage, etc.) when the agent writes no reply to the response channel within 30 seconds.

Common situations: The agent process crashed or deadlocked so the reply never arrives; a long-running query exceeds 30s; the agent's stdout reply was lost or blocked; slow machine/IO under benchmark load.

Understand the failure class

Background: Request timed out: what client-side request timeouts mean across libraries (Request timed out, TIMED_OUT, APITimeoutError) — this error's family across 39 libraries.

• Timeouts: ETIMEDOUT, deadlines, and hung requests — what actually expires when a request times out.

Related errors

• timed out during {method}
• timed out waiting for agent readiness
• ${response.Error.Message}
• Query timeout cannot be negative: " + timeoutSecs
• DBX MCP process is not available

AI-assisted analysis of t8y2/dbx@c0390bff16 (2026-09-05). Data as JSON: /api/errors/c7236e96e3e9a320. Report an issue: GitHub.