Fallback provider propagation

[lotinasa]

If I configure multiple Fallback providers and first one returns server error, 504 error for example, does it propagate to the next Fallback provider?

[ricmoo]

Can you provide the configuration? Ir if you are feeling ambitious, replicate it in the FallbackProvider test suite? :)

[lotinasa]

Question is how to achieve this, is that supported? If RPC provider returns 504 for example can I achieve to propagate request to the next RPC provider?

[ricmoo]

Is the provider throwing an error though? Like do you get a Server Error? Or is the connection just stalling? Do you have any configured priorities or weights?

[lotinasa]

Currently not throwing error, but they can through an error if that will trigger provider propagation and we are intrested what error should provider throw. We have priorities set and stall timeout.

[ricmoo]

If it isn’t throwing, and is just stalling, there is no way for the FallbackProvider to know to kick off the next request.

Once it receives a server error though, it will kick off the next. And if the stallTimeout is exceeded the next gets kicked off immediately.

It should just automatically work.

[lotinasa]

So if server error is thrown expected behavior is to propagate to the next provider?

[ricmoo]

If I understand the question, yes. The point of the FallbackProvider is to use a variety of backends, accounting for priority to get a response.