some APIs seem to have a lot of EOFErrors during non-idempotent requests, i assume due to connections being closed on the remote end. the issue affects POSTs a decent amount for some APIs (Sonos, LinkedIn, and Netatmo in particular). currently, we are retrying these requests in our client wrapper since there are no real replay risks for the particular API calls. a prior attempt to override connection.idempotent? failed, and felt off anyways. do you think this is the best course of action? we could also disable persistence altogether for connections to these APIs.
Faraday::ConnectionFailed: EOFError
protocol-http1-0.39.0/lib/protocol/http1/connection.rb:381:in 'Protocol::HTTP1::Connection#read_line'
protocol-http1-0.39.0/lib/protocol/http1/connection.rb:440:in 'Protocol::HTTP1::Connection#read_response_line'
protocol-http1-0.39.0/lib/protocol/http1/connection.rb:467:in 'Protocol::HTTP1::Connection#read_response'
async-http-0.95.0/lib/async/http/protocol/http1/response.rb:20:in 'Async::HTTP::Protocol::HTTP1::Response.read'
async-http-0.95.0/lib/async/http/protocol/http1/client.rb:83:in 'Async::HTTP::Protocol::HTTP1::Client#call'
protocol-http-0.62.0/lib/protocol/http/request.rb:87:in 'Protocol::HTTP::Request#call'
async-http-0.95.0/lib/async/http/client.rb:155:in 'Async::HTTP::Client#make_response'
async-http-0.95.0/lib/async/http/client.rb:112:in 'Async::HTTP::Client#call'
async-http-faraday-0.22.2/lib/async/http/faraday/adapter.rb:216:in 'block (2 levels) in Async::HTTP::Faraday::Adapter#perform_request'
async-http-faraday-0.22.2/lib/async/http/faraday/adapter.rb:253:in 'block in Async::HTTP::Faraday::Adapter#with_timeout'
async-2.39.0/lib/async/scheduler.rb:631:in 'Async::Scheduler#with_timeout'
async-2.39.0/lib/async/task.rb:153:in 'Async::Task#with_timeout'
async-http-faraday-0.22.2/lib/async/http/faraday/adapter.rb:252:in 'Async::HTTP::Faraday::Adapter#with_timeout'
async-http-faraday-0.22.2/lib/async/http/faraday/adapter.rb:202:in 'block in Async::HTTP::Faraday::Adapter#perform_request'
async-http-faraday-0.22.2/lib/async/http/faraday/adapter.rb:244:in 'block (2 levels) in Async::HTTP::Faraday::Adapter#with_client'
async-http-faraday-0.22.2/lib/async/http/faraday/clients.rb:107:in 'Async::HTTP::Faraday::PersistentClients#with_client'
async-http-faraday-0.22.2/lib/async/http/faraday/adapter.rb:243:in 'block in Async::HTTP::Faraday::Adapter#with_client'
async-2.39.0/lib/kernel/sync.rb:25:in 'Kernel#Sync'
async-http-faraday-0.22.2/lib/async/http/faraday/adapter.rb:233:in 'Async::HTTP::Faraday::Adapter#with_client'
async-http-faraday-0.22.2/lib/async/http/faraday/adapter.rb:176:in 'Async::HTTP::Faraday::Adapter#perform_request'
async-http-faraday-0.22.2/lib/async/http/faraday/adapter.rb:167:in 'Async::HTTP::Faraday::Adapter#call'
faraday-follow_redirects-0.5.0/lib/faraday/follow_redirects/middleware.rb:77:in 'Faraday::FollowRedirects::Middleware#perform_with_redirection'
faraday-follow_redirects-0.5.0/lib/faraday/follow_redirects/middleware.rb:65:in 'Faraday::FollowRedirects::Middleware#call'
faraday-2.14.1/lib/faraday/middleware.rb:56:in 'Faraday::Middleware#call'
faraday-2.14.1/lib/faraday/rack_builder.rb:153:in 'Faraday::RackBuilder#build_response'
faraday-2.14.1/lib/faraday/connection.rb:452:in 'Faraday::Connection#run_request'
faraday-2.14.1/lib/faraday/connection.rb:280:in 'Faraday::Connection#post'
some APIs seem to have a lot of
EOFErrors during non-idempotent requests, i assume due to connections being closed on the remote end. the issue affects POSTs a decent amount for some APIs (Sonos, LinkedIn, and Netatmo in particular). currently, we are retrying these requests in our client wrapper since there are no real replay risks for the particular API calls. a prior attempt to overrideconnection.idempotent?failed, and felt off anyways. do you think this is the best course of action? we could also disable persistence altogether for connections to these APIs.