Xray
daily rollups 1h 6h 24h 48h 3d 5d 7d 14d 30d ‹
Last 7 days▾
Ends now. Anything up to this cluster's 30-day retention.
Start and end times
Cancel
Interpreted in UTC (change it at the right of this bar). Earliest available: 2026-10-19 07:59:10.
› →|
updated 21h ago
Sample data. This database was generated by scripts/seed_dev_db.py — every service, deploy, person and figure on this page is synthetic. Names are sample.*@seed.invalid and image tags are prefixed SEED-. Nothing here happened.

X-ray, monitoring itself

Every cron job and scheduled task X-ray runs — named here on purpose, not discovered from logs. Each row's health is read from the freshness of the data the job writes; its spend is real: LLM from the call log (exact), BigQuery from X-ray's own service account priced at $6.25/TiB. Click a job for its detail, recent runs and any failures.

Local flags describe this process, not the ingester or external CronJobs. Health below is based on recorded writes, not this web replica's scheduler flag.

Notification workers · this process

Not enabled on this web process. Workers on the ingester or other processes are not described by this flag. · supervisor not running here · deadline 300 seconds

No worker attempts recorded in this process.

$4.54
TOTAL SPEND · THIS WINDOW
$4.54
LLM
$0.00
BIGQUERY
$0.65
AVERAGE PER DAY
JobRunsWhereLast runHealth LLMBigQueryTotal
Log ingest
Reads each cluster's logs (BigQuery logs_raw) and rolls up API traffic, Neo4j load and error shapes into the Cloud SQL Postgres tiers. X-ray's heaviest BigQuery reader.
scheduled
every 300s (single ingester)
master → all clusters
single ingester
not enabled here
run_scheduler=false
21h ago stale
Per-unit state (14)
UnitLast attemptLast successWindow / coverageLast error
master:api 2026-11-17 10:29:24 UTC 2026-11-17 10:29:24 UTC 2026-11-17 10:05:00 UTC → 2026-11-17 10:25:00 UTC none recorded
master:neo4j 2026-11-17 10:29:24 UTC 2026-11-17 10:29:24 UTC 2026-11-17 10:05:00 UTC → 2026-11-17 10:25:00 UTC none recorded
automation:api 2026-11-17 10:29:35 UTC 2026-11-17 10:29:35 UTC 2026-11-17 10:05:00 UTC → 2026-11-17 10:25:00 UTC none recorded
automation:neo4j 2026-11-17 10:29:35 UTC 2026-11-17 10:29:35 UTC 2026-11-17 10:05:00 UTC → 2026-11-17 10:25:00 UTC none recorded
nexus:api 2026-11-17 10:29:36 UTC 2026-11-17 10:29:36 UTC 2026-11-17 10:05:00 UTC → 2026-11-17 10:25:00 UTC none recorded
nexus:neo4j 2026-11-17 10:29:36 UTC 2026-11-17 10:29:36 UTC 2026-11-17 10:05:00 UTC → 2026-11-17 10:25:00 UTC none recorded
production:api 2026-11-17 10:30:20 UTC 2026-11-17 10:30:20 UTC 2026-11-17 10:05:00 UTC → 2026-11-17 10:25:00 UTC none recorded
production:neo4j 2026-11-17 10:30:20 UTC 2026-11-17 10:30:20 UTC 2026-11-17 10:05:00 UTC → 2026-11-17 10:25:00 UTC none recorded
delta:api 2026-11-17 10:30:56 UTC 2026-11-17 10:30:56 UTC 2026-11-17 10:05:00 UTC → 2026-11-17 10:30:00 UTC none recorded
delta:neo4j 2026-11-17 10:30:56 UTC 2026-11-17 10:30:56 UTC 2026-11-17 10:05:00 UTC → 2026-11-17 10:30:00 UTC none recorded
test27:api 2026-11-17 10:35:30 UTC 2026-11-17 10:35:30 UTC 2026-11-17 10:05:00 UTC → 2026-11-17 10:30:00 UTC none recorded
test27:neo4j 2026-11-17 10:35:30 UTC 2026-11-17 10:35:30 UTC 2026-11-17 10:05:00 UTC → 2026-11-17 10:30:00 UTC none recorded
test26:api 2026-11-17 10:40:40 UTC 2026-11-17 10:40:40 UTC 2026-11-17 10:10:00 UTC → 2026-11-17 10:35:00 UTC none recorded
test26:neo4j 2026-11-17 10:40:40 UTC 2026-11-17 10:40:40 UTC 2026-11-17 10:10:00 UTC → 2026-11-17 10:35:00 UTC none recorded
$0.000 $0.000 $0.000
Detectors
Runs the due finding detectors (budgets, outliers, job.failing, anomalies, …) after each ingest cycle and opens/resolves findings.
scheduled
after each ingest cycle
master → all clusters
single ingester
not enabled here
run_scheduler=false · detect=true · detect_live=true
2d ago stale
Per-unit state

No per-unit state available; this does not establish that the job is disabled.

$0.000 $0.000 $0.000
ArgoCD collector
Snapshots ArgoCD sync history and the ReplicaSet trail into the change feed — the deploy markers behind before/after latency comparisons.
cronjob
7 * * * * · k8s CronJob, ns default
master
external CronJob
local enablement not applicable or unknown
run_scheduler=false
— unknown
Per-unit state (10)
UnitLast attemptLast successWindow / coverageLast error
test26 2026-11-17 10:07:41 UTC 2026-11-17 10:07:44 UTC 2026-11-17 08:07:41 UTC → 2026-11-17 10:07:41 UTC
certified coverage: 2026-09-04 13:45:06 UTC → 2026-11-17 10:07:41 UTC
none recorded
test27 2026-11-17 10:07:44 UTC 2026-11-17 10:07:46 UTC 2026-11-17 08:07:41 UTC → 2026-11-17 10:07:41 UTC
certified coverage: 2026-09-04 13:45:06 UTC → 2026-11-17 10:07:41 UTC
none recorded
shared-neo4j-5-26 2026-11-17 10:07:46 UTC 2026-11-17 10:07:48 UTC 2026-11-17 08:07:41 UTC → 2026-11-17 10:07:41 UTC
certified coverage: 2026-09-04 13:45:06 UTC → 2026-11-17 10:07:41 UTC
none recorded
automation 2026-11-17 10:07:48 UTC 2026-11-17 10:07:53 UTC 2026-11-17 08:07:41 UTC → 2026-11-17 10:07:41 UTC
certified coverage: 2026-09-04 13:45:06 UTC → 2026-11-17 10:07:41 UTC
none recorded
master 2026-11-17 10:07:53 UTC 2026-11-17 10:07:59 UTC 2026-11-17 08:07:41 UTC → 2026-11-17 10:07:41 UTC
certified coverage: 2026-09-04 13:45:06 UTC → 2026-11-17 10:07:41 UTC
none recorded
nexus unknown unknown window unknown
certified coverage: 2026-09-04 13:45:06 UTC → 2026-11-04 05:07:07 UTC
unknown
delta unknown unknown window unknown
certified coverage: 2026-09-04 13:45:06 UTC → 2026-11-04 04:07:11 UTC
unknown
production unknown unknown window unknown
certified coverage: 2026-09-04 13:45:06 UTC → 2026-11-03 09:07:10 UTC
unknown
history:control-center-dev unknown unknown window unknown unknown
history:argocd-control-center2 unknown unknown window unknown unknown
$0.000 $0.000 $0.000
Skill graph
Snapshots master and production into the shared database via the API; unchanged commits reuse extraction but still report branch delivery.
cronjob
17 */6 * * * · k8s CronJob, ns default
master
external CronJob
local enablement not applicable or unknown
run_scheduler=false
— unknown
Per-unit state (2)
UnitLast attemptLast successWindow / coverageLast error
master unknown unknown window unknown unknown
production unknown unknown window unknown unknown
$0.000 $0.000 $0.000
Weaviate stats
Daily master content coverage (Neo4j versus Weaviate), with per-environment attempt/success heartbeats. Production/delta collectors are not provisioned here.
cronjob
10 2 * * * · k8s CronJob, ns tilda
master
external CronJob
local enablement not applicable or unknown
run_scheduler=false
— unknown
Per-unit state (1)
UnitLast attemptLast successWindow / coverageLast error
master unknown unknown window unknown unknown
$0.000 $0.000 $0.000
Cost ledger
Mirrors the GCP billing export into Cloud SQL Postgres so the cost pages answer in milliseconds. Reads the billing-export BigQuery dataset, not logs.
scheduled
every ~6 h (in-app, off the request path)
master
single ingester
not enabled here
run_scheduler=false · costs_enabled=false
— unknown
Per-unit state

No per-unit state available; this does not establish that the job is disabled.

$0.000 $0.000 $0.000
Baseline refresh
Rebuilds the hour-of-week baselines the anomaly detectors compare each service against (traffic, error rate, p95, Neo4j load).
scheduled
daily
master → all clusters
single ingester
not enabled here
run_scheduler=false
1d ago ok
Per-unit state

No per-unit state available; this does not establish that the job is disabled.

$0.000 $0.000 $0.000
Ask X-ray (chat)
The conversational agent: answers from the tools, proposes gated actions, reads source. The only interactive LLM spend.
llm
on demand (interactive)
master
on demand
local enablement not applicable or unknown
run_scheduler=false
— idle
Per-unit state

No per-unit state available; this does not establish that the job is disabled.

$0.000 $0.000 $0.000
Alert composer
Turns raw findings into human alert wording, narratives and rule text (Sonnet).
llm
when an alert or narrative is written
master
on demand
local enablement not applicable or unknown
run_scheduler=false
21h ago ok
Per-unit state

No per-unit state available; this does not establish that the job is disabled.

$0.085
95 calls
$0.000 $0.085
Query optimizer
Suggests cheaper Cypher/SQL shapes. Requires an LLM key; health tracks the daily master run (other clusters run weekly in maintenance).
scheduled
daily on master; weekly across ingested clusters; also on demand
master → all clusters
single ingester
not enabled here
run_scheduler=false · llm_enabled=true
— unknown
Per-unit state (1)
UnitLast attemptLast successWindow / coverageLast error
__optimizer__:master unknown unknown window unknown unknown
$4.456
58 calls
$0.000 $4.456
Watches
LLM-assisted custom watches.
llm
when a watch fires
master
on demand
local enablement not applicable or unknown
run_scheduler=false
— idle
Per-unit state

No per-unit state available; this does not establish that the job is disabled.

$0.000 $0.000 $0.000

Health: ok within cadence · late past 3× · stale past 12× · idle an on-demand job unused this window · unknown no freshness source yet.