Multiple discovery schedules active at a time

[chowdhary]

I see multiple discovery schedules active (running in parallel) in my instance. I would like to know if there will be any issue in this case and is anything going wrong with my discovery. Any suggestions or help is highly appreciated.

Thanks

Chowdhary

[tim_broberg]

Kumar, there is nothing implicitly wrong with having several discoveries running in parallel.

What you need to watch for is just getting too far behind, in which case latencies will start to rise.

If you have more probes than your mid servers can handle, you will see the latency on the ecc_queue output records start creeping up such that the time between creation and processing gets large. I saw one large customer whose ecc_queue got 40 minutes behind every day at the same time. They redistributed their discovery schedules to even out the load, which improved the situation. Also, you can add mid servers to increase your probe processing bandwidth.

If you have plenty of mids, but your nodes can't handle so many sensors, you will see the latency on the ecc_queue input records rising with much the same symptoms as they struggle to keep up with a flood of probe results coming back from the mids. Again, you can shift scheduling to even the load, or talk with support about how best to make more nodes available.

I would try to keep the latencies down below 5 minutes on both sides at all times of day.

    - Tim.

[chowdhary]

Timothy, Thanks much for sharing the necessary info. I am already seeing the latency issue on the ecc queue in my instance now. But what I didn't understand is why the same schedules are repetitively getting triggered instead of switching to next schedules as we selected 'after discovery' for all of them.

[tim_broberg]

Huh. I haven't set up a lot of schedules myself, so I'm not sure about that one. How are you expecting the schedule to be triggered? Can we get a screenshot of the schedule?

[Dave Ainsworth]

Hi Kumar,

As Tim mentions, this does sound strange. To help temporarily with this, you could change the following property to '1':

glide.discovery.max_concurrent_invocations_per_schedule

The default is 3 which means you can have 3 invocations of the same schedule happening in parallel but I don't really see a situation where anyone would want more than 1. But perhaps there are ssome cases I have not come across.

Regards,

Dave