github/openclaw
1
0
Fork 0
mirror of https://github.com/openclaw/openclaw.git synced 2026-05-08 20:08:26 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(subagents): harden announce retry guards

Browse Source
This commit is contained in:
Sebastian
2026-02-16 22:57:15 -05:00
parent f7d2e15a2e
commit 67014228cf
5 changed files with 118 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
src/agents/subagent-registry.ts
View File

@@ -1,6 +1,7 @@
import { loadConfig } from "../config/config.js";
import { callGateway } from "../gateway/call.js";
import { onAgentEvent } from "../infra/agent-events.js";
import { defaultRuntime } from "../runtime.js";
import { type DeliveryContext, normalizeDeliveryContext } from "../utils/delivery-context.js";
import { resetAnnounceQueuesForTests } from "./subagent-announce-queue.js";
import { runSubagentAnnounceFlow, type SubagentRunOutcome } from "./subagent-announce.js";
@@ -53,7 +54,16 @@ const MAX_ANNOUNCE_RETRY_COUNT = 3;
* succeeded. Guards against stale registry entries surviving gateway restarts.
*/
const ANNOUNCE_EXPIRY_MS = 5 * 60_000; // 5 minutes
// (Backoff constant removed — max-retry + expiry guards are sufficient.)
function logAnnounceGiveUp(entry: SubagentRunRecord, reason: "retry-limit" | "expiry") {
const retryCount = entry.announceRetryCount ?? 0;
const endedAgoMs =
typeof entry.endedAt === "number" ? Math.max(0, Date.now() - entry.endedAt) : undefined;
const endedAgoLabel = endedAgoMs != null ? `${Math.round(endedAgoMs / 1000)}s` : "n/a";
defaultRuntime.log(
`[warn] Subagent announce give up (${reason}) run=${entry.runId} child=${entry.childSessionKey} requester=${entry.requesterSessionKey} retries=${retryCount} endedAgo=${endedAgoLabel}`,
);
}
function persistSubagentRuns() {
try {
@@ -107,11 +117,13 @@ function resumeSubagentRun(runId: string) {
}
// Skip entries that have exhausted their retry budget or expired (#18264).
if ((entry.announceRetryCount ?? 0) >= MAX_ANNOUNCE_RETRY_COUNT) {
logAnnounceGiveUp(entry, "retry-limit");
entry.cleanupCompletedAt = Date.now();
persistSubagentRuns();
return;
}
if (typeof entry.endedAt === "number" && Date.now() - entry.endedAt > ANNOUNCE_EXPIRY_MS) {
logAnnounceGiveUp(entry, "expiry");
entry.cleanupCompletedAt = Date.now();
persistSubagentRuns();
return;
@@ -283,15 +295,17 @@ function finalizeSubagentCleanup(runId: string, cleanup: "delete" | "keep", didA
return;
}
if (!didAnnounce) {
const now = Date.now();
const retryCount = (entry.announceRetryCount ?? 0) + 1;
entry.announceRetryCount = retryCount;
entry.lastAnnounceRetryAt = Date.now();
entry.lastAnnounceRetryAt = now;
// Check if the announce has exceeded retry limits or expired (#18264).
const endedAgo = typeof entry.endedAt === "number" ? Date.now() - entry.endedAt : 0;
const endedAgo = typeof entry.endedAt === "number" ? now - entry.endedAt : 0;
if (retryCount >= MAX_ANNOUNCE_RETRY_COUNT || endedAgo > ANNOUNCE_EXPIRY_MS) {
// Give up: mark as completed to break the infinite retry loop.
entry.cleanupCompletedAt = Date.now();
logAnnounceGiveUp(entry, retryCount >= MAX_ANNOUNCE_RETRY_COUNT ? "retry-limit" : "expiry");
entry.cleanupCompletedAt = now;
persistSubagentRuns();
retryDeferredCompletedAnnounces(runId);
return;
@@ -332,6 +346,7 @@ function retryDeferredCompletedAnnounces(excludeRunId?: string) {
// Force-expire announces that have been pending too long (#18264).
const endedAgo = now - (entry.endedAt ?? now);
if (endedAgo > ANNOUNCE_EXPIRY_MS) {
logAnnounceGiveUp(entry, "expiry");
entry.cleanupCompletedAt = now;
persistSubagentRuns();
continue;