My station appears to have suffered another substantial data outage this evening. My WiFi network remained up the entire time (I was streaming a TV show), and since coming back online, the hub has not attempted to backfill. Did anyone else experience an outage at the same time (i.e. did the WF servers have another moment as they scale to keep up with the new Tempest’s coming online), or is my hub on the blink?
I’m experiencing daily data outages 2x or 3x a day ranging from 3-5 mins. My hub’s RSSI ranges from -24 thru -30. My Tempest RSSI ranges from -42 thru -46. I have Verizon FiOS as my internet provider with 1 gig down / 1 gig up.
I think the 3-5 minute gaps are expected (but unwanted) behaviour at the moment as more and more Tempest devices come online (see this thread: Dropped samples on June 15th). This longer outage may well be some of same, but it is certainly the longest I have seen so far.
That’s true. I’m noticing another issue developing…Where the web app is going down at times and not showing any metrics. Bottom line - the servers are getting hammered and timing out some from all the new units coming online.
Yep, this is probably the case. Maybe it’s 5 o clock somewhere and a bunch of Tempest have just been turned on for the first time
interesting…curious what the back bone is that runs the system. I work with AWS / AZURE… so lots of assumptions here…but if I were writing a web service of sorts to collect, I’d have my load balancers and just spin up more servers on the other side… maybe some sort of a data collection / database issue could always make things slow or back up a bit.
of course more instances == $$ for the service. Does WF make money on selling the data they collect…?
Can’t be 100% but Weatherflow uses amazon services for data. Maybe the upgrades to hubs and Tempest are messing a bit ??? Plus yes many new stations are coming online these days.
Also the devs are changing things under the hood … combine all and you have some funny things happen.
Other issue (though I hope AWS has enough to divert) hackers are wild these days … I have a dedicated server and it is flooded with intrusion pings … nuts since a few weeks and it increases each day …
perhaps you just got a firmware update during that period.
If you’re a bit tech savvy, you can capture the data directly from the hub. It broadcasts on the local network on UDP port 50222. I have a Splunk server set up to capture everything, but there are other methods of doing the same thing.
(Full disclosure: I work for Splunk.)