Datastream API is intermittently giving a Read Timed Out error

Walter Cavinaw
Walter Cavinaw Newcomer
in Datastream 
We have a number of jobs that use the Datastream API at the same time. They will each pull data (sometimes the same data) to calculate different models.

We will intermittently get a "Read Timed Out" error:

```
requests.exceptions.ReadTimeout: HTTPSConnectionPool(host='product.datastream.com', port=443): Read timed out. (read timeout=None)
```

This is very hard to replicate and only started happening in the past two weeks, but is now getting worse.

Many of our models are unable to run because of this. If we run models individually it seems to work. it seems to fail when they run at roughly the same time, or within a close interval.

Is there IP or API key throttling or rate limiting for the datastream api?

Has anything changed recently with the networking setup? (we experienced a similar problem with dsws afew months ago that was a result of datastream server config changes)

Is there any additional information I can provide to debug this?

Tagged:

Best Answer

  • Jirapongse
    Jirapongse ✭✭✭✭✭
    Answer ✓

    @Walter Cavinaw

    Thank you for reaching out to us.

    It is the HTTP timeout. The server may need more time to process the requests.

    The default HTTP timeout value of the DatastreamPy 2.0.21 library is 300 seconds while the older version is 180 seconds.

    Please check the version of the DatastreamPy library and upgrade to 2.0.21 if you are using the older version.

Categories