Intermittent timeouts from the API

Unified API
1

I have been seeing timeout errors from the API, specifically from endpoint StopPoint/Mode. The timeouts come up intermittently.

Is it possible the issue reported at the beginning of the month hasn’t been solved? API outage this monring

It has been an issue for at least the past two weeks.

2

Welcome @zuzana.svabova

I’ve not seen anything in the Cloudwatch logs that indicates the problem that I reported the other week.

Just to check … is it possible you’re being rate-limited as we are only permitted to make 500 requests a minute?

It’s useful to cache data from the API (Memcache or database table) if you can if the information you are requesting doesn’t change that often.

3

Thanks, Brian.

Interesting that you haven’t seen any errors.

We only query the API once a week for our weekly import. The import has been failing for other reasons so we had to re-run it couple times in the past two weeks but that wouldn’t be more than 2-3 requests a day.

4

Hmm.

You might want to check you used your suffixed “?app_key=” because most calls with without it but can get randomly limited.

Another thing… if it’s just a once-a-week thing, I would stick a second-long “sleep(1)” after each call, just in case…

B

5

We always include the app key so it’s not that.

Thanks for the suggestion, Brian :slight_smile:

6

hi @zuzana.svabova - the issue we had was a major incident due to the demand put on the APIs by the London Underground strikes and was resolved that morning.

What kind of level/frequency of failures are you seeing? Are they timeouts or 5xx errors? Is there a specific call or type of call failing?

thanks,
James