Since yesterday afternoon no deployment on my account can complete. Two apps are affected and I've isolated the failure to the image build queue rather than to my content. I can't clear the stuck tasks myself and I'm on the free tier, so I can't open a support ticket.
I'm leaving out account and app names here; happy to send them by DM to anyone from Posit who can act on this. The numeric ids below should be enough to locate everything.
Symptom
Every deploy to my main app (app id 16944401) fails at dispatch:
POST /v1/applications/16944401/deploy
HTTP 409 — Unable to dispatch task for application=16944401 as there are
1 tasks already in progress.
Archiving the app from the dashboard fails with the same 409, so I can't even take it out of service.
The two frozen tasks
GET /v1/tasks shows exactly two unfinished tasks on the account:
| id | action | status | app_id | created (UTC) | last update (UTC) |
|---|---|---|---|---|---|
| 1733421546 | application-deploy |
building |
16944401 | 17:54:05 | 17:54:06 |
| 1733421547 | image-build |
waiting |
— | 17:54:06 | 17:54:06 |
Both have finished: false, error: null, finished_time: null, and have not been updated since one second after they were created— many hours of zero heartbeat.
The parent application-deploy reports building, but it isn't building anything: it's blocked waiting on the child image-build, which never left the queue.
The image was never picked up
GET /v1/deployments/31825474 (the deployment referenced by task 1733421546):
image$id : 15421161
image$status : "pending"
image$active : false
image$bundle_id : 12435117
image$manifest : null
created_time : 17:54:05
updated_time : 17:54:05
status: "pending" with manifest: null and no update since creation. This is not a build that failed — it's a build that was never scheduled.
That deployment is also orphaned: the app reports deployment_id: 31824929 as the live one, so 31825474 isn't serving anything.
Isolation test — it isn't my content
The deployment immediately before this one succeeded normally with essentially the same bundle. Server log:
17:50:04 Container event from container-14250762: start
17:50:09 Listening on http://127.0.0.1:45369
That deployment (31824929) has been up and serving the whole time.
To rule out app-specific corruption I deployed the identical 6 MB bundle to a brand-new app (app id 17725977). It dispatched cleanly — no 409, task 1733437569 — and then hung in the same image-build / waiting state.
A fresh app, a fresh app id, a fresh task lane, the same result. Image builds are simply not being scheduled for this account.
What I already tried
waitForTask(1733421546)— polls indefinitely, no response.DELETE /v1/tasks/1733421546→ 400Invalid task status: buildingDELETE /v1/tasks/1733421547→ 400Invalid task status: waitingPOST /v1/tasks/<id>/canceland/abort→ 404 (endpoints don't exist)- Archiving the app from the dashboard → same 409
- Restarting R, restarting the machine, deleting the local
rsconnect/folder — no effect, as expected, since all of that is client side.
The API won't delete an unfinished task, and these tasks will never finish, so there's no self-service way out.
Possibly relevant context
The bundle is small — 6.3 MB, ~35 files, 100 R dependencies — and deployed fine all day: roughly a dozen successful deploys in the hours before this. The freeze began at 17:54, right after several deploys were launched in quick succession while a 409 was already bouncing. I suspect two overlapping deploy attempts wedged the scheduler, but that's speculation.
Ask
Could someone clear tasks 1733421546, 1733421547 and 1733437569, and release the image build queue for my account? I can provide the account name privately.
Orphan bundles left behind by the failed retries, in case they should be cleaned up too: 12435117, 12435261, 12435279, 12435386.
Happy to run any API call or provide any log that helps. Thanks.