Truncated server response: 502 Bad Gateway Error

[Sam Crockett]

Hi,

We have a number of scripts that process API pulls from various endpoints (Tasks, Jobs/Projects, etc) overnight (between the hours of 12AM - 6AM PST). While these scripts usually complete successfully, we have been receiving a truncated server response with a 502 Bad Gateway error from our API deployment address.

Here is a sample of the query that has occasionally failed:

Request failed for https://XXX.api.accelo.com/api/v0/jobs/4263/extensions/values?...

The failed request is usually related to our pull of data from the job endpoint, specifically when we query extensions (for custom fields) for each job (we loop through all of the jobs we pulled previously and then pull extension data for each job).

We would appreciate your insight on how to resolve these 502 gateway errors (or optimize our use of the API to prevent them).

Thanks!

Sam

[Matthew Hughes]

Hi Sam,

Thanks for notifying us of the errors, I will have a look into them.

Regards,
Matt

[Sam Crockett]

Hi Matt,

Following up on this thread. We are still receiving 502 and 501 server errors from the API on a semi-regular basis.

Please review and follow up when you can.

Thanks!

Sam

[Matthew Hughes]

Hi Sam,

Thanks for the follow up, my searching has not discovered much and I will pass it on to the infrastructure team (but will keep you update with here). Could you elaborate on "a semi-regular basis", would this mean several times throughout one overnight run, or a few times across several runs? Are the errors seen at other parts of the days (or do you only interact with the API through your over night scripts?).

Regards,
Matt

[Sam Crockett]

Hi Matt,

Thank you for passing this on to the infrastructure team. As for more details:

• Semi-regular basis, meaning our overnight script runs (like 'pull all tasks', 'pull all projects') may work perfectly for a few days, fail one night, work perfectly the next day, then fail again the following day.
• Our script service, Google Apps Scripts, is currently configured to terminate the script after we receive a failed request response from Accelo's API (meaning Accelo returned a 502 or similar server error).
• When these failures happen, we rerun the scripts manually, usually between 6 - 10 AM EST the following morning.
• We do occasionally see these server errors during the day time if we are testing a script or manually refreshing an overnight pull script. 
• As an example, our most recent overnight failure was on 4/12/2018 at 3:13 AM EST. A copy of our error message is below:
• Request failed for https://expertim.api.accelo.com/api/v0/tasks?_bearer_token=5RgqDwFGlFc63-zBmrzukjcS5hZ~Jj3ftkz4GuBn3y-mG1aaEeTRWMzrDUR~y3EI&_fields=id,title,status(title),date_created,date_started,date_due,date_commenced,date_completed,assignee(id,firstname,surname),manager(firstname,surname),billable,nonbillable,remaining,budgeted,task_object_budget(service_price_estimate),breadcrumbs&_limit=50&_page=428 returned code 502. Truncated server response: 502 Bad Gateway 502 Bad Gateway.
  
• For reference, based on the "_page=428" portion of the API call, you can see we made it through approximately 21,400 task records pulled before the server returned a 502.

We welcome any thoughts or feedback on how we can improve our use of Accelo's API.

Thanks,

Sam

[Kurt Wagner]

Hi Sam,

Thanks for the additional information. Next time please ensure you blank out any confidential information. e.g, instead of your actual token you may say "<bearer token>". I've deleted the token shared in your last message to prevent it being used by anyone else.

Regards,

Kurt

[Sam Crockett]

Kurt,

Thank you for catching that. I thought I had copied an abbreviated version of the full request token.

Best,

Sam

[Matthew Hughes]

Hi Sam,

An update on this, our infrastructure team have identified and taken steps to address some instability, which is potentially the cause of these errors. Please let us know if the issue arises again.

Regards,
Matt

[Sam Crockett]

Hi Matt,

Thanks for the note on this.

We experienced another timeout last night on one of our API calls for Tasks. Here is a query sample:

Request failed for https://XXX.api.accelo.com/api/v0/tasks?_bearer_token=XXX&amp;_fields=id,title,status(title),date_created,date_started,date_due,date_commenced,date_completed,assignee(id,firstname,surname),manager(firstname,surname),billable,nonbillable,remaining,budgeted,task_object_budget(service_price_estimate),breadcrumbs&amp;_limit=50&amp;_page=259 returned code 502. Truncated server response: 502 Bad Gateway 502 Bad Gateway (use muteHttpExceptions option to examine full response) (line 257, file "Code")

[Matthew Hughes]

Hi Sam,

Thanks for letting me know, I'll have a further look into it.

Regards,
Matt