Prometheus Metrics for Kubernetes CronJobs: A Complete Guide

Prometheus is the standard for Kubernetes monitoring, but CronJob observability is one of its weak spots out of the box. This guide covers everything you need to know about monitoring CronJobs with Prometheus — from the built-in metrics to custom exporters.

What kube-state-metrics Gives You

If you’re running Prometheus on Kubernetes, you almost certainly have kube-state-metrics installed. It provides these CronJob-related metrics:

CronJob Metrics

Metric	Type	Description
kube_cronjob_info	Gauge	Information about a CronJob (schedule, suspend status)
kube_cronjob_labels	Gauge	CronJob labels
kube_cronjob_created	Gauge	Creation timestamp
kube_cronjob_next_schedule_time	Gauge	Next scheduled run (Unix timestamp)
kube_cronjob_status_active	Gauge	Number of currently running jobs
kube_cronjob_status_last_schedule_time	Gauge	Last scheduled time (not last success!)
kube_cronjob_spec_suspend	Gauge	Whether the CronJob is suspended

Job Metrics

Metric	Type	Description
kube_job_info	Gauge	Information about a Job
kube_job_owner	Gauge	Job’s owner reference (links to CronJob)
kube_job_status_succeeded	Gauge	Number of succeeded pods
kube_job_status_failed	Gauge	Number of failed pods
kube_job_status_active	Gauge	Number of active pods
kube_job_complete	Gauge	Whether the job completed
kube_job_status_completion_time	Gauge	Completion timestamp
kube_job_status_start_time	Gauge	Start timestamp

Useful PromQL Queries

Last Execution Status per CronJob

This requires joining CronJob and Job metrics, which is the biggest pain point:

# Check if the most recent job for each CronJob succeeded
kube_job_status_succeeded{job_name=~".*"}
  * on(job_name) group_left(owner_name)
  kube_job_owner{owner_name=~".*", owner_kind="CronJob"}

CronJob Execution Duration

# Duration of the last completed job
(kube_job_status_completion_time - kube_job_status_start_time)
  * on(job_name) group_left(owner_name)
  kube_job_owner{owner_kind="CronJob"}

Detect Missed Schedules

# CronJobs where the last schedule time is older than expected
time() - kube_cronjob_status_last_schedule_time > 86400

This is approximate — it compares current time to last schedule, but doesn’t account for the CronJob’s actual schedule interval.

Failed CronJobs in the Last Hour

# Jobs owned by CronJobs that have failed pods
kube_job_status_failed > 0
  unless on(job_name) (kube_job_status_succeeded > 0)
  * on(job_name) group_left(owner_name)
  kube_job_owner{owner_kind="CronJob"}

The Gaps in kube-state-metrics

While kube-state-metrics provides raw data, it has significant limitations for CronJob monitoring:

1. No direct CronJob success/failure metric — you have to join across Job and CronJob metrics
2. No execution counter — you can’t easily count total executions over time
3. No duration tracking — calculating duration requires joining start and completion times
4. No true missed schedule detection — you can approximate it, but it’s unreliable
5. High cardinality — Job names include timestamps (e.g., nightly-backup-28486760), which creates new time series constantly

Purpose-Built CronJob Metrics with Varax Monitor

Varax Monitor was built to fill exactly these gaps. It watches CronJob executions directly and exports clean, purpose-built metrics:

Metric	Type	Description
cronjob_last_execution_status	Gauge	1=success, 0=failure per CronJob
cronjob_last_execution_duration_seconds	Gauge	Duration of last execution
cronjob_execution_total	Counter	Total executions by success/failure
cronjob_missed_schedules_total	Counter	Total missed schedules
cronjob_next_schedule_time	Gauge	Next expected execution time
cronjob_is_suspended	Gauge	Suspension status

These metrics are labeled by CronJob name and namespace — no complex joins required.

Example: Alert on Any Failed CronJob

With kube-state-metrics:

# Complex multi-metric join
kube_job_status_failed > 0
  unless on(job_name) (kube_job_status_succeeded > 0)
  * on(job_name) group_left(owner_name)
  kube_job_owner{owner_kind="CronJob"}

With Varax Monitor:

# Direct, simple query
cronjob_last_execution_status == 0

Example: CronJobs Running Longer Than Usual

With kube-state-metrics:

# Requires multiple joins and history comparison
(kube_job_status_completion_time - kube_job_status_start_time)
  * on(job_name) group_left(owner_name)
  kube_job_owner{owner_kind="CronJob"}
  > 300

With Varax Monitor:

cronjob_last_execution_duration_seconds > 300

Setting Up a Complete Dashboard

Whether you use kube-state-metrics alone or Varax Monitor, your Grafana dashboard should include:

1. Overview table — all CronJobs with last status, last run time, and next scheduled time
2. Failure timeline — when and which CronJobs have failed
3. Duration trends — execution time over the last 7 days per CronJob
4. Missed schedule alerts — any CronJobs that didn’t fire when expected
5. Active jobs — currently running CronJob executions

Varax Monitor includes a pre-built Grafana dashboard with all of these panels. Install it with one command and customize as needed.

Getting Started

If you want the simplest path to CronJob observability:

helm repo add varaxlabs https://charts.varax.io
helm install varax-monitor varaxlabs/varax-monitor

It’s free, open-source (Apache 2.0), and deploys in under 60 seconds. Read the full quickstart.