Unannounced downtime - isochrones endpoint

Hi, thanks for making this useful + wonderful service freely available! We are using it to make a resource for helping people understand COVID-19 data in their immediate neighorhood.

The following endpoint is returning 503 errors for me consistently, even though statuspage says that it’s up and running. My API key has not been rate limited.


Wanted to check in to see if this was intended downtime, or if this might be unexpected. Thanks for any help you can provide!

HI, I’ve just restarted the server so hopefully it will be back up in half an hour. It seems people have been making very heavy requests against the car profile the last hour and so it killed it

Seems to be working now, thank you!

Is there a health check endpoint that we can visit without an authorization token? If you need help with monitoring endpoint uptime, I would be interested in helping to set up a monitoring/alerting solution.

I’ve just updated the status page so that should hopefully show when individual profiles are down in the future (all runs through StatusCake).

And thanks for the offer - we have some internal monitoring checks in place (I actually restarted the servers before I read your message yesterday) - I just missed the first couple of emails saying that the server was down. We are currently looking into how we could automatically restart the servers when they go down, but that’s a little tricky as quite a few things need to be thought about to get that working properly :slight_smile:, and even then, each profile would still need half an hour to reload.