Something wrong with servers or feeds?

It makes one week that I have this issue. On Android, the application crashes after trying to refresh for a long time

seemed better for a bit but still having issues today

George, if it were one of the servers behind the load balancer, the load balancer would see. Those timeouts would be provided by the load balancer (HAProxy in this case). If you get your browserā€™s timeout, then you canā€™t even connect ot the www server, which is a bigger problem and something I canā€™t fix.

Been horrible for me for the past couple of days. Ā Have been unable to get feeds to load in the Android app, had issues with items not getting marked as readā€¦ Everything that the folks above are mentioning has been happening to me too.

Getting the same behavior in the Android app on wifi or on mobile data, as well as with FF and Chrome on the desktop (Windows 7, 8, and 10). Page appears to load but the wheel of loading never stops and you see a permanent message like the attached image. This is at work as well as at home, so different internet providers.

Been going on nearly all weekend for me.

Ā 

I added a few new front-end servers, but nothing looked wrong. Iā€™m fairly certain thereā€™s something upstream thatā€™s causing this.Ā 

Anybody who is on the android app, are you on the latest beta?

Hello Samuel, something like this helps? My NewsBlur today, the feed is the link called sapo

Whatever you did, there are no longer any delays. Everything is loading lightening fast now.

BTW, my Android version is 4.5.0

Iā€™m on 4.5.0 on my android, everything seems fast and working now on both my android and one of my windows comptuers

Same here - everything is working MUCH better today than it has been for the past week

My android app version is 4.5.0.

I just signed up to become a beta tester, so hopefully in a few hours Iā€™ll have the beta version to try out.

Ditto ā€“ it appears that things are back to normal (for me at least ā€“ web and iOS use, multiple ISPs) over the past couple of hours. Iā€™d be curious to know what changed (if it was an ā€˜upstreamā€™ issue, then Samuel didnā€™t change any code, but presumably DigitalOceanā€™s providers fixed something).

So hereā€™s a graph that will confirm the issues and the timeline.Ā 

This shows the past 7 days. It measures the time it takes to fetch and parse feeds, sampling from the past five minutes. Those two lines, page and feed_fetch, are 100% network I/O. (feed_process is separated out). On average, NewsBlur will fetch and parse a feed, its original page, update the icon, and count unreads in under 3 seconds, ~7 million times a day.

Looks like there were 4.5 days of network latency. That would explain it. I immediately logged a support request with Digital Ocean last week, but they didnā€™t find anything.

This is now fixed for me after you added more servers. Thanks

Things seem good here now. Ā Thanks.

http://applite.eu/