Cronhq

Many teams end up running a small server just to call a URL at 2 a.m. Cronhq is a hosted scheduler that does this instead: a developer points it at a webhook, and it fires the call on schedule, retries on failure, and sends an alert when something breaks and again when it recovers. The site frames it as cron jobs that actually run, and describes itself as written in Rust on top of Postgres, currently version 0.1.

Creating and syncing jobs

Jobs can come from a REST API, from the dashboard, or from a cronhq.yaml file kept in the repository and reconciled by the command-line tool, so schedules sit in version control. Each job is defined by a time zone, the address to call and a cron expression with five fields. The dashboard has a command palette where typing a phrase such as "every weekday at 9am" returns a valid cron expression with a plain-English preview.

Exactly-once runs and retries

A worker claims each run with a Postgres lock, which the vendor says prevents two workers from firing the same slot. Failed calls are retried after a delay set per job, and every attempt is logged with status code, duration and a shortened copy of the response body. Requests are signed with HMAC-SHA256 under a secret that belongs to the job, so the receiving service can verify the call came from Cronhq, and the secret can be rotated without downtime.

Heartbeat monitors

Heartbeat monitors work the other way round. A job pings Cronhq on every run, and if a ping is late, the monitor flips to down and an alert is sent. Email or Slack alerts fire after the third failure inside an hour and once more when the job recovers.

Plans, and what could not be checked

The service has a free tier and two paid tiers that differ in the number of jobs, the minimum interval and how long history is kept, and annual billing is arranged by email. The pricing page also mentions a free container for self-hosting, licensed under MIT, but the GitHub repository linked from that page returned an error when checked, and the npm package named in the install command was not found on the registry, so neither the license nor self-hosting could be confirmed. No terms, privacy policy or company name appear on the site, and it does not say where data is stored.

Cronhq is a hosted scheduler aimed at developers who need HTTP calls on a timetable, without running cron servers of their own. Jobs are set up with a REST API, the dashboard, or a cronhq.yaml file that a command-line tool syncs, using a five-field cron expression plus a time zone and the URL to call.

What Cronhq offers

• Single claimed runs: a Postgres lock is meant to stop two workers firing the same scheduled run.
• Retries and logs: failed calls are retried with a per-job delay, and each attempt is logged with its status code, duration and a truncated response body.
• Signed requests: each request carries an HMAC-SHA256 signature made with a per-job secret.
• Heartbeat monitors: a job pings Cronhq on each run, and Cronhq alerts when a ping is late.

Who it works with

Developers scheduling webhook or HTTP calls who want alerts by email or Slack.

Key facts

• Alerts: sent on the third failure within an hour, and again on recovery.
• Build: written in Rust on Postgres, at version 0.1.
• Self-hosting: the pricing page offers a self-hosted container under the MIT licence, but the GitHub repository it links to returned not found, and the npm package in its install command is not on the npm registry.