Monitoring on Satellite Services

Question

Hi All,

Not sure if this is the right place to post this: first time poster.

I'm having some difficulty with monitoring of certain devices 3 satellite services in service at the moment: From our datacenter, cross-connect to different datacenter in another state, from this datacenter to the site via satellite.

Most of it works good: customer has connection and full access can be arranged. Though we are having one issue, the service will appear as "down" at random times for random amounts of time (can be 2 minutes, can be 6 hours): it is accessible and we can ping the service during these times. It is causing a major annoyance to our technical team with all the false alerts.

Are there any tips or solutions that others have implemented to solve this? Solarwinds support suggested changing the polling criteria, though I don't believe that would help it (as it goes down for hours at times)

Thanks,

James

mjalden1 · Answer

I have remote sites with same issue. I also run sat com links and they appear down when not but it also does the same for regular nodes. Its best to check your poller. I have multiple pollers and when the issue occurs I move that service/node to another poller and it shows up. I then proceed to reboot the bad poller. I have seen this issue when system team pushes updates that have DotNet inside them like cumulative updates. The update will get installed all services are still up but have issue where poller doesn't receive snmp updates until its rebooted. After reboot i move services back and check polling.

Jfrazier · Answer

unfortunately that is the joys of satellite based services unless you are lucky enough to have low earth orbit birds but those are expensive to use.  Anything like Hughesnet is going to have the higher ping latency.

As a note, years, okay about 2 decades ago when I worked at Radioshack, we were trying to work with Satellite for store connectivity in rural areas where DSL, cable, etc were not available. We had over 5000 stores at the time.  We had the same issues using HP's Network Node Manager. There was not really any way around it and the latency kept causing VPN to drop, etc.  In the end we scrapped satellite as a viable solution.  On paper it looks workable but in real life it just is really not there.  Wish I had better news...

James.Rich · Answer

Average response time is between 611ms - 675ms.

The ICMP and SNMP Timeout is set to 2500ms - when it goes down on solarwinds, we can ping the device with the normal latency (around 600ms) perfectly fine. SNMP retries is set to 5 - but Solarwinds is set to poll every minute, though it stays down for hours.

I honestly believe the issue is elsewhere and not in solarwinds.