Never retry any of the failed requests

[danigutierrezayuso]

We've integrated the Snowplow sdk on a project where we don't want to retry any of the failed requests.

We've configured the tracker so it doesn't retry any of the failed requests with HTTP error status codes, using customRetryForStatusCodes on the EmitterConfiguration. But we're still seeing that all the requests that failed because of a timeout or any other connection error are being saved to the database and the tracker tries to send them again.

It would be great to have a way to define on that EmitterConfiguration that we don't want to retry any of the failed requests. Something like:

EmitterConfiguration()
    .retryFailedRequests(false)

Changing the implementation of the shouldRetry function on RequestResult class returning this value directly when it's set.

We want to avoid having traffic peaks with a lot of events being sent at the same time, for example after fixing an issue on the snowplow server. And right now there is no way to configure the tracker to do that.

Does this change make sense to you?

Thanks!!

[matus-tomlein]

Hi @danigutierrezayuso, I think this makes sense! Adding a retryFailedRequests option to the EmitterConfiguration would be a good way to go about it. I'll add this to our backlog, but also very happy to receive a contributions for this!

In case you are looking for a temporary workaround – one approach would be to create a custom NetworkConnection implementation that always returns a successful status code no matter whether the request failed or not. You can then pass this into NetworkConfiguration: https://snowplow.github.io/snowplow-ios-tracker/documentation/snowplowtracker/networkconfiguration