Cloud Scheduler reports status.code=-1 despite HTTP 200 attempts

Environment

I am troubleshooting an isolated, non-production Google Cloud setup.

- Cloud Scheduler region: “europe-west1”

- Cloud Run Job region: “europe-southwest1”

- Schedule: “*/15 * * * *”

- Time zone: “Etc/UTC”

- Target: HTTP POST to the Cloud Run v2 Jobs “:run” API, authenticated with OAuth.

Observations on 17 October 2026 (UTC)

During five scheduled slots at approximately 20:30, 20:45, 21:00, 21:15 and 21:30 UTC, Cloud Scheduler logs showed “AttemptStarted” and “AttemptFinished” entries. Each finished attempt reported HTTP 200.

Five Cloud Run Job executions were also observed. Each showed “succeededCount=1” and a completion timestamp. Cloud Run audit logs contained five “Jobs.RunJob” events referencing those executions.

However, some execution creation/start timestamps precede the corresponding Scheduler “AttemptStarted” log timestamps. We could not establish exact causal correlation using the available audit output.

State after pausing the Scheduler

The subsequent “gcloud scheduler jobs describe” output showed:

- “state: PAUSED”

- “status.code: -1”

- “scheduleTime” absent

- “lastAttemptTime” absent

The Console also displayed “Has not run yet”, despite the successful attempts in the logs.

Questions

1. Is “status.code=-1” expected for a paused Scheduler job after attempts returned HTTP 200?

2. Why might “scheduleTime” and “lastAttemptTime” be absent when attempt logs exist?

3. What does “userUpdateTime” represent in this situation?

4. What is the recommended way to correlate Scheduler attempts with Cloud Run v2 “Jobs.RunJob” audit events and execution resources?

5. Does “Has not run yet” indicate expected behavior, stale Console state, or a discrepancy requiring investigation?

I am not assuming a service defect. I would like to understand the expected API and Console behavior and identify the minimum read-only evidence needed.

Both Scheduler jobs remain paused. No production resources, IAM permissions, canonical collector, models or historical data were changed.

Hey, welcome to the community! We’ve seen your question, and while we wait for a community member to chime in and share their thoughts, check out the Cloud Scheduler Documentation to learn more.

We appreciate your participation in the community!

Following up with a more precise description of the issue.

The concern is not that Cloud Scheduler failed to receive HTTP 200 responses. During five scheduled slots, the logs showed five AttemptStarted/AttemptFinished pairs with HTTP 200, five successful Cloud Run Job executions, and five Jobs.RunJob audit events.

However, the available evidence does not uniquely correlate each Scheduler attempt with each execution.

After both Scheduler jobs were paused, the resource showed state=PAUSED, status.code=-1, and no scheduleTime or lastAttemptTime.

Could someone clarify the documented semantics of these fields for a paused job and recommend the read-only evidence needed to establish the correlation?

We are not claiming a product defect. Both jobs remain paused, and no further executions or infrastructure changes are being performed.

I would not read the -1 as a failure of those attempts, since your own logs show AttemptFinished with HTTP 200 and five succeeded executions. To me it reads as the describe output having no recorded last-attempt status for the paused job, which also fits the missing scheduleTime and lastAttemptTime and the Console’s ‘Has not run yet’ text. On userUpdateTime, my understanding is that it tracks the last update to the job itself, such as pausing it, rather than the last run, so it will not help you correlate attempts. For correlation, I would match by time window and counts, five attempts, five executions and five audit events in the same 15-minute slots, since the Scheduler attempt logs and the Cloud Run audit logs do not share an execution identifier, and I would not treat small timestamp ordering differences across two regions as a mismatch on their own.

Thanks, this is helpful and gives us a plausible explanation.

Before treating this as resolved, could you point me to a Google Cloud reference or documented behavior confirming that “status.code = -1” and a missing “lastAttemptTime” are expected for a paused job?

The REST reference describes “status” as the response from the target for the last attempted execution, and “lastAttemptTime” as the time the last attempt started. I could not find documentation explaining why these fields would be absent after pausing.

We will keep the jobs paused while we establish the expected behavior. The five attempts and five executions are consistent with successful operation, but we still cannot prove a one-to-one causal correlation from the available evidence alone.

I’ve been digging into the same paused-job case on my own project since I posted, and the REST reference I found does define the status and lastAttemptTime fields for a job, which helped me understand why a paused job shows no recent attempt. What I couldn’t find is a page that actually names -1 as the expected code for a paused job, so I don’t want to claim the docs say that when they don’t. From what I saw in my own paused jobs, the missing lastAttemptTime lined up with the job simply never having run since it was paused, rather than an error I needed to fix. I’d still keep yours paused as you planned and compare the output again after a manual run, since that gave me the cleanest before-and-after check.

Thanks for checking this in your own project. I understand your explanation is a plausible hypothesis, but we still need to establish the documented behavior for our case.

The REST reference defines “status”, “scheduleTime” and “lastAttemptTime”, but does not explain whether “status.code = -1” and the absence of these timestamp fields are expected after pausing a job that previously completed successful attempts.

Could someone with authoritative Google Cloud knowledge clarify this behavior or point us to the relevant documentation?

We also still need to understand whether there is a documented, read-only method to correlate individual Cloud Scheduler attempts with Cloud Run v2 “Jobs.RunJob” audit events and execution resources when no common request or operation identifier has been established.

We will not perform a manual run or resume either job during this investigation. Both jobs remain paused, and no infrastructure changes are being made.

Thank you for any clarification or official documentation you can provide.