Hi,
Using socket-apollo-link with an Apollo client, if the Absinthe GraphQL resolver on our server throws an error, our client and server seem to get stuck in a cycle where the client action is not aborted but keeps retrying and failing. The raw websocket cycle looks something like:
client: send action
server: ["9","9","__absinthe__:control","phx_error",{}]
client: ["21","21","__absinthe__:control","phx_join",{}]
server: ["21","21","__absinthe__:control","phx_reply",{"response":{},"status":"ok"}]
client: resubscribe to subscriptions
back to top...
How would you recommend handling this on the client? Ideally, the client should have a way to associate a server crash with a permanent error. Is there an exposed method we can use to abort the action (aside from disconnecting the socket)?
Hi,
Using socket-apollo-link with an Apollo client, if the Absinthe GraphQL resolver on our server throws an error, our client and server seem to get stuck in a cycle where the client action is not aborted but keeps retrying and failing. The raw websocket cycle looks something like:
How would you recommend handling this on the client? Ideally, the client should have a way to associate a server crash with a permanent error. Is there an exposed method we can use to abort the action (aside from disconnecting the socket)?