Repository navigation
Zen GLM 5.3 Flash: repeated ~605s first-generation delays and unknown finishes #50628
Description
Zen / GLM 5.3 Flash: repeated approximately 605-second first-generation delays and unknown finishes
We use native OpenCode 1.18.23 on Linux amd64 with opencode/glm-5.3-flash,
reasoning max. Please investigate the request path/provider routing for these
sessions. All times below are UTC. No prompts, customer information, credentials
or reasoning content are included.
On 22 September 2026 we observed three long intervals before the first recorded
reasoning event. These are client observations, not proof of provider queue time:
| Session | Assistant request record | First reasoning event | Delay |
|---|---|---|---|
| ses_f37d79ef9ffelEzl7A5IVqwuMA | 08:08:05.113 | 08:18:10.846 | 605.733s |
| ses_f37dbfd60ffeSm2TVnZ8jAwl0q | 08:03:39.629 | 08:13:42.433 | 602.804s |
| ses_f37dbfd60ffeSm2TVnZ8jAwl0q | 08:15:22.252 | 08:25:26.677 | 604.425s |
The first session subsequently produced a complete answer. The latter two turns
were in one outer session. Existing investigation reported zero retries; our
saved wrapper receipts confirm no wrapper retries. These records alone do not
prove whether an upstream gateway retried internally. Client memory had ample
headroom and the daemon did not restart during this interval.
Separately, the following 21 September native invocations ended with unknown finish
and no visible answer. Each wrapper receipt records one native invocation and zero
wrapper retries. A cleanup SIGKILL followed the incomplete response; it is not
evidence of an OOM or of the original upstream failure. Usage was unavailable,
not proven zero. We do not claim these share a root cause with the delays above.
| Session | Started UTC | Elapsed seconds |
|---|---|---|
| ses_f3c0adab2ffei7vd3V2Y2ygphQ | 2026-09-21T12:33:35.508801+00:00 | 102.511 |
| ses_f3c085008ffeyH7pBlUdN06dQz | 2026-09-21T12:36:22.429373+00:00 | 242.056 |
| ses_f3c03942affewv8LNXx3kQR9OD | 2026-09-21T12:41:32.430222+00:00 | 271.811 |
| ses_f3bfe34bbffeNbfbsWOLeLeXFU | 2026-09-21T12:47:24.904347+00:00 | 253.263 |
| ses_f3bfd9b75ffeeW6pbRdAkWLK6q | 2026-09-21T12:48:04.135005+00:00 | 93.468 |
| ses_f3bf94921ffedErcoSV4EYrriX | 2026-09-21T12:52:46.653119+00:00 | 199.044 |
| ses_f3be5d646ffeSfQyAE6aAXcOHS | 2026-09-21T13:14:01.533272+00:00 | 381.634 |
| ses_f3bd4fed9ffe7SE511EWgJkZn9 | 2026-09-21T13:32:25.641846+00:00 | 217.541 |
| ses_f3bcbb268ffelKbIyQhK32bZhT | 2026-09-21T13:42:35.082769+00:00 | 74.881 |
| ses_f3bb99299ffeNN87n0h4xkKLv3 | 2026-09-21T14:02:22.633589+00:00 | 299.393 |
| ses_f3bb773b3ffe4GTr4HLSC0pR5s | 2026-09-21T14:04:41.542269+00:00 | 146.888 |
| ses_f3ba2b5e3ffeCrVh4oliOko4rJ | 2026-09-21T14:27:21.140273+00:00 | 239.664 |
| ses_f3b9e7d6dffeUtDx7IBt86o5Ml | 2026-09-21T14:31:57.713122+00:00 | 109.928 |
| ses_f3b9db4fcffeJh7eaJkz1IStNc | 2026-09-21T14:32:48.873121+00:00 | 222.848 |
| ses_f3b999e59ffeZnn6qFrEEOJO1G | 2026-09-21T14:37:16.762067+00:00 | 160.732 |
| ses_f3b9586f6ffehRI7r9cxtrky5Q | 2026-09-21T14:41:45.102169+00:00 | 190.508 |
| ses_f3b93669bffe03jVGpuhReb7SC | 2026-09-21T14:44:04.376565+00:00 | 262.778 |
| ses_f3b8a48dcffe0J6NAEHJtzYWYU | 2026-09-21T14:54:01.567863+00:00 | 269.783 |
Can you correlate these sessions and timestamps with Zen's upstream routing,
queueing, retries, idle deadlines and terminal stream events? Is an approximately
600-second boundary expected anywhere in that path? Which request-id/header is
reliable for future support correlation? Our previous Bun response-header logging
returned no headers and cannot establish HTTP/request-id provenance retroactively.
No additional provider runs have been started to reproduce this support report.
Activity
github-actions commentedon Sep 22, 2026
This issue doesn't fully meet our contributing guidelines.
What needs to be fixed:
Please edit this issue to address the above within 2 hours, or it will be automatically closed.
This issue might be a duplicate of existing issues. Please check:
If you believe this was flagged incorrectly, please let a maintainer know.
github-actions commentedon Sep 22, 2026
This issue has been automatically closed because it was not updated to meet our contributing guidelines within the 2-hour window.
Feel free to open a new issue that follows our issue templates.