A lot has been added to Cronjob.de in recent weeks. You create cronjobs with AI, read errors in plain words, get outages reported where you work, and keep an eye on jobs on your own server, too. From the first call to monitoring, you have everything with one provider.
All jobs are running on schedule.
A set of instructions or a single sentence becomes a finished cronjob.
Paste the instructions from your host, shop or CMS. The AI picks out the address, method and schedule, asks if something is missing, and fills in the form. You check it and save.
Just write when it should run and you get the schedule along with the next dates. For an exception such as “except Sundays between midnight and 10 am”, we name two cronjobs that split the times exactly.
Cronjobs for a campaign or a period switch themselves off at the time you choose. Schedule and log are kept, and the AI also understands “until 15 November”.
See what is running, understand what is stuck, and pause everything cleanly for maintenance.
A timeline shows what is running right now and what comes next, and refreshes itself every minute. From Plus you see the result of every call up to 24 hours back and the success rate. Below it, a band shows every day of the last 30 days, with Unlimited of the last 90 days, plus the runtimes.
From Plus, every error is spelled out in words in the list, dashboard and call log, with advice on what you can do. If Cloudflare or another protection service stops the call with a security check, we recognise it and tell you how to set up an exception there.
Calls and errors per day at a glance, plus every error period with start, end and cause. So even weeks later you can see when it got stuck and why.
For work on your server, pause everything or single folders with one click. When you end it, exactly what ran before starts again, and if you like, it ends by itself after 1, 4, 12 or 24 hours.
Charts of the duration and size of the last 200 calls show when a script slowly gets slower, long before it hits the time limit.
You hear about a problem before your customers notice. Also for jobs on your own server.
If a cronjob fails, we report it by e-mail, in Slack, Microsoft Teams or Discord, or as a webhook to your own system. You connect Slack with one click; for Teams and Discord you paste the webhook address. You get a message when the outage starts and the all-clear as soon as it runs again.
Notifications are newly arranged: Where messages go, What they cover, How often we report, and what was Reported last.
We check every minute whether your site really works, and recognise error pages despite “200 OK”, empty pages and a missing add-to-cart button. You enter the address, the AI suggests the checks, and a second server confirms every outage.
For jobs on your own server, in n8n or in Laravel: your job checks in with us with one line after every run. If the check-in is missing, we alert you, and when it comes back, we send the all-clear. The graphic above shows how it fits together.
The code to add is available for Linux, PHP, Python, Node.js, n8n and many more. For Lovable, Bolt, Cursor and Claude Code you get a ready-made prompt for the AI. With /$? at the end, the line also reports the exit code: anything other than 0 counts as an error.
For cronjobs we call: one line at the very end of your script reports that it ran all the way through. So you also hear about an abort in the middle of the script. You switch it on in the cronjob under “Success check”, or in the list with the finish flag.
In the cronjob’s new “Success check” tab you set which text must appear in your script’s response, such as “Import complete”, and which must not. If the response differs, the call counts as an error. From Plus we also recognise typical error messages such as “Fatal error” without your own texts.
Once a day we check how long the certificate of your addresses is still valid, and let you know in good time: 14 days before it expires and once more 3 days before. You find it under “Monitoring” next to the servers.
Work together in the same account and manage cronjobs in a chat.
Colleagues get their own login with their own e-mail address and password, as admin or member. A log shows who created, changed or deleted which cronjob, and agencies switch between their clients’ accounts in the menu at the top.
Connect Claude, ChatGPT, Cursor or another assistant to your account and simply ask what went wrong last night. On request, the assistant also creates cronjobs and heartbeats and switches on “Monitor the end”. You decide whether it may only read or also create and change.