CLOUD · OPERATIONS

Operations teams: scheduled jobs with no server to maintain

No servers

to patch or watch for the reports and follow-ups that run every day

Challenge

Several of our clients relied on small scheduled tasks: a queue of text messages, a nightly follow-up, a monthly finance report.

These ran on an office computer or an aging server. They failed silently when that machine had a bad day, and the first sign was a customer who had not heard back.

Approach

We rebuilt the tasks as serverless functions in the cloud, which run only when there is work to do.

  • Rewrite each task as a small function kept in version control.
  • Trigger it on a schedule or when work is waiting in a queue.
  • Retry automatically when an outside system is unavailable.
  • Report failures to a person the same day.
  • Store the files each job produces so any result can be traced.

Result

The jobs now run whether or not anyone is in the office.

  • No server to patch for routine work.
  • Failures reported, not discovered.
  • Cost that follows actual use.

Our advice

Look for the tasks that run on a machine under someone's desk. They are usually cheap to move and expensive to lose.

Related service: Cloud computing.

Let’s talk about where your technology should be in a year.