Blog · Monitoring
5 critical tasks to monitor with Mortemain (and why)
Dozens of jobs run automatically on your servers. The question is who watches them. Here are the five whose silent failure hurts most, and the one line each one needs.
When a scheduled job stops without a word, the fallout is real: lost data, blocked workflows, an expired certificate taking a site down. None of it errors, so your dashboard stays green. A Mortemain check-in turns each one into an alert. Start with these.
1. Backups
A backup that isn't running means unprotected data, and you find out during a restore, too late. Ping on success at the end of the job:
rsync -avz /data/ backup:/backups/data/ && curl -fsS -m 10 https://ping.mortemain.com/your-uuid
Per tool: pg_dump · restic · Borg · all backups
2. Cron jobs
Log cleanup, cache refreshes, scheduled emails. Drop one by accident or let it die quietly and nobody notices. Add the ping to the crontab line:
0 3 * * * /usr/local/bin/report.sh && curl -fsS -m 10 https://ping.mortemain.com/your-uuid
See the cron guide and the 3-step walkthrough.
3. CI/CD pipelines
A pipeline that fails blocks deploys. One that never triggers is invisible to your CI tool. A final step that always runs closes the gap:
- name: Notify Mortemain if: always() # even on failure run: curl -fsS -m 10 https://ping.mortemain.com/your-uuid
Per tool: GitHub Actions · GitLab CI · Jenkins
4. Data syncs (ETL, Airflow)
A sync that stops leaves your systems quietly out of step. Ping at the end of the script:
import requests requests.get("https://ping.mortemain.com/your-uuid", timeout=10)
Per tool: Python · Airflow · Prefect · Dagster
5. Certificate renewal
An expired certificate takes the whole site down, and classic monitors don't always catch a failed renewal. Report success, or /fail if the renewal errors:
certbot renew --quiet && curl -fsS -m 10 https://ping.mortemain.com/your-uuid \ || curl -fsS -m 10 https://ping.mortemain.com/your-uuid/fail
Which to monitor first
Prioritise the jobs that:
- have no redundancy, so nothing takes over if they stop;
- are business-critical (backups, payments, billing);
- run unattended, with no human watching (cron, CI/CD);
- can fail silently, with no logs and no error.
Pick one, add its check-in, then simulate a failure to confirm the alert lands. Try Mortemain free (20 checks, no credit card).