regsvc (tilda) image pull failing for at least 117 min — 1 pod
Pod cannot start: regsvc (tilda) image pull failing for at least 117 min — 1 pod. Facts: namespace tilda, workload regsvc, why image pull failing, pending_minutes 117, since 1794904982.
What X-ray is prepared to defend about test27 right now — each one a claim
with its numbers, the evidence to check it, and what normal looks like. Detectors last ran
never.
?A finding is opened by a detector — a statistical test over the stored rollups, or a
live check of a source — and stays ONE row while the condition persists: it updates,
escalates, and resolves itself fifteen minutes after the detector stops seeing it.
Acknowledge to say you have it; snooze to hide it for a while; nothing here is a
threshold somebody configured.
Every card is a claim X-ray can back with numbers you can click through to. Findings rank by severity first, then by an impact score (severity band + the subject's share of the cluster's traffic + how long it has been open), so the top of the page is the thing to look at first.
What this page will not tell you
Pod cannot start: regsvc (tilda) image pull failing for at least 117 min — 1 pod. Facts: namespace tilda, workload regsvc, why image pull failing, pending_minutes 117, since 1794904982.
The background job pubsub.onboarding-test27-regsvc-sub on test27 failed 19 runs recently (528 failure lines; lines within one run count once), detected via its ERROR-severity log lines. A job that logs its failures below ERROR severity — a WARNING carrying an error field — is invisible to the novel-shape detector and to the page's own failed count; this detector reads the error field itself, which is why the finding now appears. The sample failure is in the facts.
The background job studioai.unknown on test27 failed 5 runs recently (532 failure lines; lines within one run count once), detected via its ERROR-severity log lines. A job that logs its failures below ERROR severity — a WARNING carrying an error field — is invisible to the novel-shape detector and to the page's own failed count; this detector reads the error field itself, which is why the finding now appears. The sample failure is in the facts.
Pod cannot start: rag-agent (tilda) image pull failing for at least 118 min — 1 pod. Facts: namespace tilda, workload rag-agent, why image pull failing, pending_minutes 118, since 1794904947.
jss-server in tilda on test27 has restarted in place 9 times since 2026-11-17 10:05 UTC (9 in the last 30 minutes, 9 in 24 hours), most recently at 2026-11-17 10:15 UTC. Kubernetes' account of it: 11 crash events, 3 sched events. Last event: BackOff — “Back-off restarting failed container jss-server in pod jss-server-5968bc86d8-pptmg_tilda(fda1a1d7-5f21-42bd-b73d-7d92a2cf2383)”. It is running image spreadsheet-agent-edits-d3202c0. Containers restarting: jss-server ×9. The workload carried 0.0% of the cluster's API calls in the last day. Restarts are counted from the container restart counter, so a rollout or a spot preemption does not count; a crash does. It resolves on its own after thirty quiet minutes.
Autoscaler blind: HorizontalPodAutoscaler/tmfsync-hpa (tilda) cannot read its scaling metric — 7× in 2 h. Facts: namespace tilda, object HorizontalPodAutoscaler/tmfsync-hpa, failures_2h 7, last_message No recommendation.
A workplansvc query shape on test27 has p95 1.4 s in the second half of the window against 488 ms in the first (3.0×, 223 executions). The shape reads: “MATCH (wp:WorkPlanInstance)-[:WORKPLAN_INSTANCE_TO_MILESTONE]->(:WorkPlanMilestone)-[:REL_WORKPLAN_INSTANCE_MILESTONE_TO…”. Slower can mean more data, a lost index, or lock contention — the shape page has the CPU/wait split. Shapes whose p95 stays under a second are never reported: below that nobody notices, whatever the ratio.
onboardingsvc on test27 has served no API calls for 40 minutes (last call 2026-11-17 05:00 UTC); at this hour of the week its baseline — 5 weeks of history — expects about 132 calls an hour, 66 in the window checked. It carried 10.6% of the cluster's calls over the last week. Silence is not measured downtime: the pods may be up and unused, or crashed, or cut off upstream — Reliability and Pods say which. It resolves on its own once the service serves again.
4xx spike (403/401/404…): studiosvc GET /api/studio/triggers: permission denied — 84 of 84 calls (100 %; 401×84) in the last 15 min — 5× its usual 16.3. Facts: endpoint GET /api/studio/triggers, service studiosvc, calls 84, c4xx 84, share_pct 100.0, c401 84.
A docsvc query shape on test27 has p95 1.6 s in the second half of the window against 69 ms in the first (23.9×, 1,858 executions). The shape reads: “MATCH(a:Organization{deltimestamp:$a_deltimestamp,alias:$a_alias}) RETURN a…”. Slower can mean more data, a lost index, or lock contention — the shape page has the CPU/wait split. Shapes whose p95 stays under a second are never reported: below that nobody notices, whatever the ratio.