Backend server migration

[hozn]

Hi folks –

Just a quick note that we’ve moved backend server to a new system. This has happened in stages, but I just put in the redirect rule to actually forward people from the [old] https://freezingsaddles.com domain to the [new] https://freezingsaddles.org domain. (For those that are curious, the backend is now split into 3 different application containers: a webhook listener, a sync engine, and the website.)

You probably won’t notice any differences in the site (though I think Jon pushed up a few fixes that hadn’t yet been deployed to the .com version, but should be on the new one), but this is a new architecture for how we get rides from Strava that will be far less taxing on the API. The old model was generating hundreds of errors per day, so my inbox is quite relieved by this move.

Also, you should notice that your rides show up almost immediately on the leaderboards, as we’re using the Strava webhooks system. Webhooks isn’t perfect, though, so if you crop a ride or make other minor changes (e.g. classify it as a “commute”) that doesn’t get pushed down to us as a change. However, we do resync everyone’s rides once a day (we divide the athletes up into 24 bins and sync one bin every hour) just in case something gets missed.

We’ll also add a resync button/option in the UI. (There’s one for photos now, but I think it’s broken anyway.)

I’m pretty sure that there are going to be bugs in the new system. Feel free to PM me or, better, email hans at velum dot net. Or even better open a ticket at https://github.com/freezingsaddles/freezing-web/issues . And for anyone that has some Python knowledge or wants to learn, we have lots and lots of open tickets that need addressing. Now that this rearchitecture is done, I will start going through the various leaderboard requests, etc.

Topic

[kwarkentien]

Thanks for all that you and Jon do behind the scenes to keep our silly little challenge up and running and awesome!!!!!

Sent from my iPhone using Tapatalk

[jrenaut]

This one was almost entirely Hans. Quite an impressive restructuring of the entire website in an extremely short amount of time.

[chuxtr]

I am getting a Failed to Open, Can’t Connect to Server when I try going to freezingsaddles.org

[CBGanimal]

@chuxtr 174070 wrote:

I am getting a Failed to Open, Can’t Connect to Server when I try going to freezingsaddles.org

Me too…was looking forward to the new site but can’t get it to load…although I have to say a lot of sites I’ve been going to gives me the same error!

Sent from my iPad using Tapatalk

[peterw_diy]

Yep, unable to connect. Hans, I recommend Uptime Robot for decent free monitoring with no strings attached. You can point it at anything that responds with a status 200 item or even a standard HTTP auth challenge (it treats an auth challenge as evidence of Uptime). So you could even point it at your webhook service…

Good luck with the repair, and thanks as always for all your work!

[jrenaut]

I don’t know why we’d need an uptime service when we clearly already have an effective one.

[hozn]

Sorry, guys. They rebooted the digital ocean droplet and apparently the nginx container wasn’t set to start on boot.

Yeah, I use uptime robot for other things so will either use that or one of the metrics backends to send alerts if/when this happens again.

[streetsmarts]

@kwarkentien 174064 wrote:

Thanks for all that you and Jon do behind the scenes to keep our silly little challenge up and running and awesome!!!!!

Sent from my iPhone using Tapatalk

Agreed. Thanks yall!÷

Sent from my SM-G900V using Tapatalk

[chuxtr]

FYI, my ride today isn’t showing. And when I try to log in on the Freezing Saddles page, i.e., connect with my Strava, I get this:

{“message”:”Bad Request”,”errors”:[{“resource”:”Application”,”field”:”redirect_uri”,”code”:”invalid”}]}

[anomad]

This thread reminds me of our last version release during the open application period at work…

[david.helms570@gmail.com]

The site is responding but many rides logged today on Strava haven’t been tallied on the BAFS page.

@hozn 174080 wrote:

Sorry, guys. They rebooted the digital ocean droplet and apparently the nginx container wasn’t set to start on boot.

Yeah, I use uptime robot for other things so will either use that or one of the metrics backends to send alerts if/when this happens again.

[hozn]

@david.helms570@gmail.com 174085 wrote:

The site is responding but many rides logged today on Strava haven’t been tallied on the BAFS page.

Yeah, that was probably because they came in while the site was offline (so it would have caught up after a day). In any event, this should have been fixed preemptively earlier this evening. Let me know if you still see any issues.

[hozn]

@chuxtr 174083 wrote:

FYI, my ride today isn’t showing. And when I try to log in on the Freezing Saddles page, i.e., connect with my Strava, I get this:

{“message”:”Bad Request”,”errors”:[{“resource”:”Application”,”field”:”redirect_uri”,”code”:”invalid”}]}

Thanks for the heads up; I needed to let Strava know that the domain had changed. (This is done now.)

[awitt]

My am ride was on the FS site immediately. Thanks!

[chuxtr]

This from one of my team members who has forum phobia. She rode yesterday morning, her ride is on Strava, but as of late last night it wasn’t on the FS leaderboard.

“I’m not sure why but my VERY COLD ride today, though on Strava, still hasn’t loaded onto the leaderboard. I’m concerned. Is there anything we should do other than wait?”

[Author]

Replies