intermittent nginx 502 bad gateway errors

I am receiving intermittent 502 nginx “bad gateway” errors, and even when the page loads, the feeds in it do not.


I’m not seeing bad gateway errors, but Newsblur is really struggling today. Most of the time I get no content after the blue bar has finished doing its thing.

Yeah I’m getting the same as Daniel - no nginx errors, but no content either.

Same here. Feeds are not loading - it says “everything is on its way” but nothing loads.

I’m having similar problems, with the added bonus that sometimes the API throws me 404 when I try to request a story.

Same. Been getting these issues for about 2 days now.

I had problems with feeds loading VERY slowly all day yesterday. Right now it looks like things are a bit better and I’m not seeing error messages.

I’m working on getting things back to normal. My graphs are showing everything being normal, so I’m having a hard time figuring out what it is. Additionally, I’m on vacation, so I have very limited connectivity and time.

I just pushed a small code update that fixes the Oh No errors, and hopefully added some better logging. I won’t be able to really fix the problem until tomorrow. But it looks like things are stable again (and you should confirm here that it is or isn’t, so I know).

I can confirm that it loaded without 502 errors, and after a few page refreshes the “Everything is on its way” went away. Clicking on a feed in the tree takes about 45 seconds after the JS progress-bar in the bottom panel is “complete” before content appears.

Things always break on the weekend and holidays, eh? :slight_smile:

I just restarted just about every service I have and now things are looking OK.

MUCH better! Not taking so long to load the feeds now. Thanks!

Is everything back to 100% normal speeds? Watch that graph of average load times on the front page. That is my only indication that everything is going well (combined with a whole bunch of other graphs, but that one is my go to). If that looks OK, but things are taking forever, please post here.

speed has definitely improved compared to this morning.

It’s working normally now. Thanks!

I concur, it seems the magic of the reboot button has cured all the ills.

Thanks a bunch for taking time away from your vacation to look into this.

What worries me is that this happened at all. I have a few systems that are allowed to fail, and I have automated processes in place to correct those failures. But what happened today was not acceptable and should never have happened in the first place. I’ll be looking into the root cause, as well as reliving some of the strain of the system by finishing one of my performance branches this week.