Hi Fred,

• What are SDL's recommendations regarding the OAuth2 authentication? In particular, do you recommend the use of the Refresh token (as it seems there may be reliability issues)?
  • Yes we recommend using the Refresh token. I’m not aware of any stability issues – have they reported these issues to us? Happy to look into any problem…
• Is it possible to obtain multiple Access tokens from the same public IP?
  • Yes, this is possible.
• Is it possible to maintain concurrent connections? How many?
  • No “connection” limit is enforced, but a future release may introduce API call throttling. We have the facility to control this on a per user/integration level – reaching the limit would result in a 409 response until the timeframe passes.
  • However, multiple identical calls to the same endpoint are often “serialized” internally – e.g. 10 connections all simultaneously asking for a list of projects will be processed sequentially (per user), and this can sometimes extend the time required for each response – internal caching helps where possible (we’ve seen this with the previous Viator integration, for example, and have improved things internally, but it springs to mind here…)

Thanks
Kris

Hi Fred,

• What are SDL's recommendations regarding the OAuth2 authentication? In particular, do you recommend the use of the Refresh token (as it seems there may be reliability issues)?
  • Yes we recommend using the Refresh token. I’m not aware of any stability issues – have they reported these issues to us? Happy to look into any problem…
• Is it possible to obtain multiple Access tokens from the same public IP?
  • Yes, this is possible.
• Is it possible to maintain concurrent connections? How many?
  • No “connection” limit is enforced, but a future release may introduce API call throttling. We have the facility to control this on a per user/integration level – reaching the limit would result in a 409 response until the timeframe passes.
  • However, multiple identical calls to the same endpoint are often “serialized” internally – e.g. 10 connections all simultaneously asking for a list of projects will be processed sequentially (per user), and this can sometimes extend the time required for each response – internal caching helps where possible (we’ve seen this with the previous Viator integration, for example, and have improved things internally, but it springs to mind here…)

Thanks
Kris