Naming conventions for a scheduler system

[lahma]

There are generic rules for messaging systems in the documentation, but I'm looking for ideas and verification for naming scheduling systems. Easiest to think about *nix CRON system where there's scheduled tasks running and there's both load (how many running at the time and how fast each job runs).

There's earlier thread, but there was little output from chosen names and patterns.

My specific scenario is Quartz.NET (a .NET scheduling system, ported from its Java counterpart). At high level:

• jobs written in code (a .NET type implementing a callback interface)
• multiple triggers can target jobs, they can be CRON-based or say timer of every 5 seconds and some repeat count (or indefinite)
• there can be multiple nodes (processes and/or machines) processing job invocation when trigger's fire time is reached

I'm working on built-in support for exposing traces and metrics here. So thus far naming and data is as follows, feedback and ideas appreciated!

💡 I'd also like to raise a question whether such scheduling/job invocation concept should be formalized as there are many such libraries and services out there in the wild.

Activity names

• Quartz.Job.Execute - job is being executed, the actual thing the library does
• Quartz.Job.Vetoed - job execution was vetoed for some reason, very rare

Activity tags

• scheduler.name - the name of scheduler, e.g. "Warehouse"
• scheduler.id - instance id, e.g. "warehouse-node-02"
• fire.instance.id - unique fire id, think primary key in database
• trigger.group - trigger group, e.g. "Maintenance"
• trigger.name - name "HourlyCheck"
• job.type - the .NET type in question hosting the running code, e.g. "WareHouseCleanupJob", bound to code
• job.group - job group, e.g. "Maintenance"
• job.name - job's name "MaintenanceJob"

Meters

These feel like a more generic concept than Quartz!

• scheduling.quartz.execute - counter, total number job jobs that have been executed
• scheduling.quartz.execute.errors - counter, error that have been raised during job execution
• scheduling.quartz.execute.active, counter - currently running job count
• scheduling.quartz.execute.duration, histogram - how long job execution took

Meters have the tags

• trigger.group
• trigger.name
• job.group
• job.name
• scheduling.quartz.exception_type, only if exception occurred