How to Get Alerted When a Workflow Stops Succeeding
Open an incident when any workflow starts failing, and close it automatically once it succeeds again.
Overview
A workflow that quietly starts failing is easy to miss until a customer notices. In Spojit, a Stopped succeeding monitor watches every workflow you choose and opens an incident for each one that has failed a set number of times in a row since it last succeeded. When the workflow succeeds again, the incident resolves on its own and everyone who was told hears that too.
Each workflow gets its own incident, so one broken workflow does not hide another. Turn on suggested next steps and Miraxa looks into each failure and suggests what to fix.
Before You Start
- You need permission to manage monitors.
- Watching all workflows or whole folders needs permission to view every workflow in the workspace.
Steps
- Open Control Center and go to the Monitors tab.
- Under Recommended for this workspace, choose Set up on Workflows stop succeeding. It opens in the editor with sensible settings.
- Select the what it watches part of the sentence and choose which workflows: All workflows, Chosen workflows or Folders. Turn on Only workflows that are active to skip drafts.
- Set how many failures in a row should open an incident. The recommended monitor uses 3.
- Select who to tell and pick a Slack or Teams contact, or Email me.
- Choose the level. Critical suits workflows your customers depend on.
- Under More options, turn on Ask Miraxa for suggested next steps if you want Miraxa to investigate each failure.
- Choose Test it now to see which workflows would alert right now, then Create monitor.
Other Run-Health Monitors
- Failed runs - a workflow failed a number of times over a period, even with successes in between.
- Error rate - the share of a workflow's finished runs that failed passes your limit. Judge only with at least a number of runs so one failure on a quiet workflow does not count as 100%.
- Stuck runs - a run is still going after the time you set. Runs waiting for an approval do not count.
Tips
- Use Leave out for workflows that are expected to fail sometimes, such as ones that retry against a flaky supplier.
- Pair this monitor with a Stuck runs monitor. A workflow that hangs never fails, so it never stops succeeding.
- Choose Folders rather than chosen workflows when a team owns a folder. New workflows added to the folder are watched automatically.
Common Pitfalls
- Setting the failures in a row to 1 on workflows that hit temporary errors. A single failure that succeeds on the next run then opens and resolves an incident each time.
- Watching inactive workflows. Turn on Only workflows that are active so drafts and paused workflows do not alert.
- Expecting this monitor to catch workflows that never run. It only sees runs; use an Ask Miraxa monitor to alert when a workflow has not run for a while.
Related Articles
- Creating a Monitor
- How to Send Monitor Alerts to Slack or Microsoft Teams
- Applying Suggested Fixes and Using Auto-Apply
- Using Miraxa to Debug and Explain Failed Runs
- Understanding Execution Logs
- Monitoring Workflow Executions