Monitoring with status hooks

Fisher doesn't have any monitoring or reporting solution built-in, for a simple reason: there are countless ways you could do that, and integrating all of them would be a daunting task. Instead, Fisher provides status hooks, a way to build the integration yourself in a simple way.

Check out the tutorial for an hands-on introduction.

Status hooks execution

Status hooks are executed when an event happens inside of Fisher, allowing you to react to it. The following events are supported:

  • job-completed: a job completed without any error
  • job-failed: a job failed to execute, probably due to an error

Status hooks are executed in the scheduler along with the normal jobs, but with a priority of 1000. This means they will be executed before any other job, but you can override this behavior by giving the most important scripts an higher priority.

Creating status hooks

To create a status hook, you just need to create a script that uses the Status provider:

## Fisher-Status: {"events": ["job-failed"], "scripts": ["hook1.sh"]}

The status hook is configured with a configuration comment, and supports the following keys:

  • events: the list of events you want to catch
  • scripts: execute the status hook only for these hooks (optional)

Execution environment

Status hooks are executed with the following environment variables:

  • FISHER_STATUS_EVENT: the name of the current event
  • FISHER_STATUS_SCRIPT_NAME: the name of the script that triggered the event
  • FISHER_STATUS_SUCCESS: 0 if the script failed, or 1 if it completed
  • FISHER_STATUS_EXIT_CODE: the script exit code (if it wasn't killed)
  • FISHER_STATUS_SIGNAL: the signal that killed the script (if it was killed)
  • FISHER_STATUS_STDOUT: path to the file containing the stdout of the script
  • FISHER_STATUS_STDERR: path to the file containing the stderr of the script