New user since 2 months , Location: The Netherlands
I notice regular outages of the whole network , Most certain the problem is not on my side
It seems they allways occur around/after midnight (European time).
Question-1 : Why ?
Question-2 : Why is there no notification anywhere ? most networks give some notification.
Question-3 : Why isn’t there a single member inquiring about this ?
New user since 2 months , Location: The Netherlands
first, welcome to the RS community and thanks for becoming a part of the world’s largest seismic network. your questions reminded me that newcomers are likely not aware of the history of the Raspberry Shake seismometer and our data services. this history is, in fact, pertinent to your queries.
we very recently celebrated our 7th birthday, after another year of continued growth in terms of the total number of shakes online at any given moment. what started out as a good idea, initially funded through a kickstarter project, has grown to become the largest global seismic network to have ever existed. this did not happen by accident: without exception, each and every individual who works with the RS company is exceedingly motivated and dedicated to provide the best product and service possible, whether the end-user is a hobbyist, a professional, or an educator. all of us here believe deeply in what we do, that we are making a difference, that because for the first time a high-performing seismometer is now available to the masses at low-cost, being able to participate in “citizen science / seismology” is now possible. this is an extraordinary achievement of which we are all extremely proud.
on to your first question:
being that we are a network that is permanently growing means that, from time to time, upgrades to the server infrastructure must be carried out, all the while making sure that the network remains fully functional (24/7/36) to the greatest extent possible. as it happens, this is exceedingly difficult, (but for a person like me, actually a lot of fun), resulting in some down time where access to the data is concerned.
it is unclear from your question what exactly you mean by “outages”, but at the moment, this is the situation where the network and the data are concerned:
- for the near future, it is possible that the real-time data feed will be temporarily suspended (< 2 minutes)
- for the near future, it is possible that data requests from the FDSNWS service will be temporarily suspended (< 2 minutes)
however, it is very important to note that the current solutions in place are meant to prioritize one thing:
- that no data is lost, i.e., that it is accessible at a later time for retrieval through the FDSNWS service
this is clear by viewing the helicorder page of the dataview app.
what this means, then, is that it is entirely possible that the real-time data feed will be potentially compromised during the time that we are diligently working to upgrade our back-end systems.
so when you say “outages”, if you mean the real-time feed, then yes, that can happen. but if you mean permanent data loss, then this mostly does not happen, where maximizing data completeness will always be our priority over all others.
also, to be clearly stated:
- all data services provided by RS servers are free, both the service and the data itself. while we at RS do the best we can to keep these services up and available 24/7, we cannot, and will not, provide any guarantee that this will always be true.
- RS data feeds should absolutely NOT be used for any type of follow-on critical services, such as alerts or any other function an end-user can think of. RS is not an early warning system, nor a service to be used for any purpose where 100% reliability is required.
as regards notifications of network status, there is a twitter channel dedicated to this. however, notifications are typically generated only when there is a serious disruption which may last for a longer amount of time than usual. but when the real-time data feed is suspended for < 2 minutes, when no data is ultimately lost, no notification is generated.
as for why no one else is inquiring about this, i can’t say for sure, but it’s entirely possible that the end-users who do closely monitor the data feeds know and understand what i’ve explained above, knowing that all of us here are constantly working, in the background, to make the network more resilient, scalable, and accessible to everyone in the world, today and for the future.
re: not scoring any points with you
i had intended to reply to you yesterday, my apologies for not being able to get to that before today. as it turned out, contrary to your observation that i was ostrach-izing my head, i was busy, working on making sure that no data is lost, while also making upgrades to the system to handle the ever-increasing load; we are indeed a victim of our own success, good for us, and good for all the end-users: because of this success we will continue to exist.
i hope that i have clarified the situation for you, and that you have a better understanding of what we are, and what we aren’t. if what we are does not match up with your expectations of what we should be, we will regret losing you: a citizen scientist participating in an experiment that has never before been tried in the seismology domain (which will surely have hiccups from time to time). but, as opposed to just turning it off, letting it collect dust, perhaps a local school might be interested in taking it on as part of their science curriculum?
Thank you for your extensive reply.
May i suggest a real-time / live “Status-Indicator” ?
something to avoid having to think “What is going on ?”
A Graph, Icon(s) or Text : like: “Status: We are updating, May take 5 to 40 Mins)”
I am a little bit surprised. Isn’t there an advertising slogan: “Professional Solutions”?