TencentCloud/TencentDB-Agent-Memory · error

${TAG} ${path} fetch failed: ${(err as Error).message}

Error message

${TAG} ${path} fetch failed: ${(err as Error).message}

What it means

MetaClient.fetch wraps the metadata-service HTTP POST; any exception thrown by fetch() itself (network error, DNS failure, timeout via AbortSignal.timeout) is rethrown with a prefixed message naming the path and underlying cause. This distinguishes transport-level failures from HTTP status and envelope errors, which have their own messages.

Source

Thrown at MemoryProxy/src/meta/client.ts:542

    const headers: Record<string, string> = {
      "Authorization": `Bearer ${this.serviceToken}`,
      "x-tdai-service-id": this.serviceId,
      "Content-Type": "application/json",
    };
    if (this.userKey) {
      headers["x-tdai-user-key"] = this.userKey;
    }

    let resp: Response;
    try {
      resp = await this.fetcher(url, {
        method: "POST",
        headers,
        body: JSON.stringify(body),
        signal: AbortSignal.timeout(this.defaultTimeoutMs),
      });
    } catch (err) {
      throw new Error(`${TAG} ${path} fetch failed: ${(err as Error).message}`);
    }

    if (!resp.ok) {
      let detail = "";
      try {
        detail = await resp.text();
      } catch { /* ignore */ }
      console.log(
        `[wb-debug] metadata-client req path=${path} status=${resp.status} url=${url} serviceId=${this.serviceId} userKey.len=${this.userKey?.length ?? 0} userKey.prefix=${(this.userKey ?? "").slice(0, 20)} serviceToken.len=${this.serviceToken?.length ?? 0} body.len=${detail.length} body.head=${detail.slice(0, 300)}`,
      );
      throw new Error(`${TAG} ${path} HTTP ${resp.status}${detail ? `: ${detail.slice(0, 200)}` : ""}`);
    }

    const env = (await resp.json()) as CoreEnvelope<T>;

    if (env.code !== 0) {
      throw new Error(`${TAG} ${path} envelope error ${env.code}: ${env.message ?? ""}`);
    }

View on GitHub (pinned to 3efcd317b8)

Solutions

1. Confirm the metadata service URL is correct and reachable from the runtime (curl the endpoint)
2. Increase defaultTimeoutMs if timeouts are the cause (AbortError in message)
3. Check network egress/proxy/firewall rules for the deployment environment
4. Retry with backoff for transient network errors; inspect the inner error message after 'fetch failed:' for the root cause

Example fix

// before
new MetaClient({ ..., defaultTimeoutMs: 3000 });
// after
new MetaClient({ ..., defaultTimeoutMs: 15000 }); // tolerate slow metadata service

// precheck connectivity
const reachable = await fetch(serviceUrl, { method: "HEAD", signal: AbortSignal.timeout(3000) }).then(() => true).catch(() => false);
if (!reachable) throw new Error("metadata service unreachable before request");

try {
  await client.createTask(body);
} catch (e) {
  if (/fetch failed:.*(TimeoutError|AbortError|ECONNRESET|ENOTFOUND)/i.test(e.message)) {
    await backoffRetry(() => client.createTask(body), 3);
    return;
  }
  throw e;
}

When it happens

Trigger: createTask/updateTask/appendParticipationLog calls fetch and the request rejects before a response is returned: connection refused/reset, DNS resolution failure, TLS error, or defaultTimeoutMs elapsing (AbortError/TimeoutError).

Common situations: Metadata service down or wrong URL configured; corporate proxy/firewall blocking egress; timeout too low for slow networks; service DNS not resolvable inside the deployment environment.

Related errors

• KernelFetchError(504 on timeout, 502 otherwise) with dynamic
• [instance-config] Config source returned empty VDB config fo
• [skill-agent-queue] tasks-mutex wait timeout for ${key}
• Embedding API returned unexpected format: missing 'data' arr
• task_not_found

AI-assisted analysis of TencentCloud/TencentDB-Agent-Memory@3efcd317b8 (2026-09-01). Data as JSON: /api/errors/35847651bb6951d4. Report an issue: GitHub.