New in 0.10 · Kubernetes and servers · every edition

SSL certificate expiry and cron job monitoring: Know before a certificate expires or a job stops.

WatchLog finds every TLS certificate your servers and clusters serve, every systemd timer, cron job and Kubernetes CronJob, and alerts when one is about to expire, fails, misses its run or runs far too long. Nothing to list, nothing to wrap in a ping.

Applications › Certificates and Scheduled jobs

Certificates

  • api.acme.iohaproxy.service on edge-1 :443ExpiringIts renewal job certbot.timer is failing5 days left
  • mail.acme.iopostfix.service on mail-1 :465ProblemNot seen renewing yet22 days left
  • shop.example.comIngress shop/storefrontWorkingRenews itself; last time 32 days ago57 days left

Scheduled jobs

  • backup.timersystemd timer · worker-1 · Every day at 03:00FailingFailed 3 times in a row. Last good run: 16 NovFailed, exit code 2
  • export.phpcron · app-1 · Every hour at :15SlowRunning 4 times longer than usualRunning for 41 min
  • nightly-reportCronJob in shop · Every day at 02:00WorkingNothingSucceeded

The alertbackup.timer failed 3 times in a row on worker-1exit code 2 · last good run 3 d ago · its runs are on the job’s page

An example. Names are illustrative.

Found by the agents you already run

Certificates, timers, cron jobs and CronJobs

  • Certificates on Linux servers

    The agent reads the certificate every listening TCP port presents, STARTTLS included for SMTP, IMAP, POP3, PostgreSQL and MySQL, and asks for each name in the server’s Let’s Encrypt, acme.sh and Caddy stores, so every virtual host is found.

  • Certificates on Kubernetes

    The cluster-agent reads each TLS Service and each Ingress host through its controller, the API server’s own certificate included. No Secret is opened.

  • systemd timers and cron jobs

    Timers from systemd itself, enabled or not. cron jobs from the crontabs, matched with the processes cron really starts, so each run has its duration and exit code.

  • Kubernetes CronJobs

    Each Job a CronJob creates is a run; a failed Job’s pod gives the exit code. The CronJob’s time zone is honoured.

No traffic is decrypted, no private key or Kubernetes Secret is read, and a cron job’s command line never leaves the server.

What you get on day one

  • Alerts before it’s an outage

    A warning 14 days before a certificate expires and critical at 3 days; a job that fails, misses a run or runs far longer than usual. On from the first day, in every edition.

  • Renewals you can trust

    Every certificate seen under the same names is kept, so WatchLog says when one last renewed and how early, and names the certbot or acme.sh job that renews it when that job fails.

  • Missed is really missed

    A random delay, a catch-up after a reboot or a daylight saving change is never “late”, and a run due while the server was off is told apart from one that didn’t start.

  • Only a name leaves the server

    A cron job is named from its command (pg_dump, artisan schedule:run). Its command line, arguments, environment and user never leave the server, so a password in a crontab stays there.

  • Why it broke

    Probable cause names an expired certificate where failing connections arrive, and the failing timer behind a service that failed.

  • Ask your AI

    The MCP tools list_certificates and list_scheduled_jobs answer “Which certificates expire this month?” and “Did last night’s backup run?”.

Part of one product

The same dashboard, alerts and probable cause for your Kubernetes clusters, Docker hosts and Linux servers.

KubernetesDockerLinux servers

Questions

How does SSL certificate expiry monitoring work here?

The agent connects to each TLS port on the server, or each TLS Service and Ingress host in Kubernetes, and reads the certificate it presents every 10 minutes, as a browser would. No list of hosts to keep, no traffic decrypted and no private key read.

How does cron job monitoring work without changing my crontab?

The agent reads the crontabs and sees each command cron starts with a small eBPF program, so it knows each run’s start, end and exit code. Nothing to wrap in a ping, no heartbeat URL.

Does it monitor systemd timers?

Yes: every timer, enabled or disabled, with its schedule, random delay and catch-up setting, and each run of the unit it starts, with its exit code.

What isn’t covered?

A virtual host whose name is in no certificate store, anacron’s and at’s own jobs, cron jobs inside containers that run their own cron, cron’s CRON_TZ lines, certificates on UDP, and on Kubernetes Ingresses without TLS and the Gateway API.

Do I need a new agent?

Agent 0.10 or later, and on Kubernetes the chart’s 0.10 permission to list Ingresses. The tabs name the servers whose agent is older.

What does it cost?

It is in every edition. Self-hosted is free and open source (AGPL-3.0). WatchLog Cloud Free covers 5 nodes or servers; Pro is $3 per node or server a month, $15 minimum.

No more surprise expiries, no more silent cron jobs.

In every edition, free and open source to self-host. Cloud Free covers 5 nodes or servers; Pro is $3 per node or server a month, $15 minimum.