Hi, my ISP (AS8422) has - at the time of writing - a major core network outage. None of its customers have internet access.
I tought this would be an interesting use case for the ripe atlas and monitored the situation, creating a measurement to regularily ping all available probs in this AS: RIPE Atlas - RIPE Network Coordination Centre
This worked well until 2025-02-12 12:33:03 UTC, when the last probe reported a status. After that it just stopped on its own (I scheduled the measurement to end on one day later). I waited for a bit and at 2025-02-12 14:15:21 UTC decided to manually stop the measurement and tried to create a new one.
But creating a new measurement for this AS was not possible and it gave me an error: “There was a problem with your request Details: 1. Your selected ASN is not covered by our network.”
So my questions would be:
Why did my initial measurement suddenly stop?
Why could I not create a new measurement?
I speculate the answer to the second question could be that it was because there were no probes online in this AS at the time. Which would be unfortunate, because I thought of the ripe atlas as a tool that is able to monitor exactly this kind of situation.
Since your original measurement used probes located in AS8422, and those probes all went offline, no results were able to be collected from them. When you tried to start a new measurement also with probes from AS8422, the system refused for the same reason – there were no available probes in AS8422.
Unfortunately if all probes in a given AS are disconnected there is nothing we can do to get results from them. What you could try in future is to use one or more pingable targets in the AS8422 announced address space, but specify the source probes as fully random, or at least in networks outside of AS8422.
I hope this helps, please let me know if I’ve misunderstood your issue or if you have any further questions.
Why would the measurement stop tho? As you can see in the measurement linked above, the latest result had some probes being pingable.
Is is the case that if in one measurement-iteration no probe replies, the measurement gets automatically stopped?
I understand that if a probe is not reachable, no ping can be executed on it and the probe of course can’t return a result. But “probe didn’t reply” is a valuable result as well in my opinion, especially when monitoring network issues.
Maybe my use case is out of scope and not an intended use case for atlas, I just wanted to give some feedback as to how I tried to use it and what issues I ran into.
Thank you for the tipp to add some probes outside of the actual networking I’m interested in, I will keep that in mind.
Please keep in mind that using probes as targets is unlikely to succeed or be reliable (NATs, firewalls, downtimes, …). We suggest targeting anchors where that’s applicable - in this case it isn’t… but we would welcome one of course!
As for your original measurement: probes cannot report results while they are disconnected (naturally) but they keep doing the measurement in the meantime and submit all results when they connect again. I think this happened in your case as well. Eventually they heard your request to stop, and by 16:00 (UTC) all of them did so.
