TaskQueueOutstandingCount server metric

Is this metric reasonable to use as a “task queue backlog”? I’ve seen a few other threads here that suggest I should be monitoring schedule_to_start latency and that there isn’t a task queue backlog metric, but they are all fairly old threads and I’m not sure if TaskQueueOutstandingCount is new since then.

You should use enhanced describe task queue api, see docs here

Thank you!

Since these new metrics aren’t available via the temporal check in datadog – we’d need to write our own watchdog to pull task queue metrics via the enhanced describe task queue api and emit them to datadog.

In the meantime, is TaskQueueOutstandingCount meaningful to continue using with our alerts?