To help support temporary errors that can occur in the capnp rpc
calls, a wrapper is introduced to inspect the error conditions and
allow for retrying within a short window.
Combines the tunnelrpc and quic/schema capnp files into the same module.
To help reduce future issues with capnp id generation, capnpids are
provided in the capnp files from the existing capnp struct ids generated
in the go files.
Reduces the overall interface of the Capnp methods to the rest of
the code by providing an interface that will handle the quic protocol
selection.
Introduces a new `rpc-timeout` config that will allow all of the
SessionManager and ConfigurationManager RPC requests to have a timeout.
The timeout for these values is set to 5 seconds as non of these operations
for the managers should take a long time to complete.
Removed the RPC-specific logger as it never provided good debugging value
as the RPC method names were not visible in the logs.
If cloudflared was unable to register the UDP session with the
edge, the socket would be left open to be eventually closed by the
OS, or garbage collected by the runtime. Considering that either of
these closes happened significantly after some delay, it was causing
cloudflared to hold open file descriptors longer than usual if continuously
unable to register sessions.
## Summary
We discovered that we were being impacted by a bug in quic-go,
that could create deadlocks and not close connections.
This commit bumps quic-go to the version that contains the fix
to prevent that from happening.
Before this commit the commands that listed tunnels and tunnel routes would be limited to 1000 results by the server.
Now, the commands will call the endpoints until the result set is exhausted. This can take a long time if there are
thousands of pages available, since each request is executed synchronously.
From a user's perspective, nothing changes.
## Summary:
In order to properly monitor what is happening with the new write timeouts that we introduced
in TUN-8244 we need proper logging. Right now we were logging write timeouts when the safe
stream was being closed which didn't make sense because it was miss leading, so this commit
prevents that by adding a flag that allows us to know whether we are closing the stream or not.
## Summary
To avoid having to verbose logs we need to only log when an
actual issue occurred. Therefore, we will be skipping any error
logging if the write timeout is caused by no network activity
which just means that nothing is being sent through the stream.
This commit makes the remote diagnostics enabled by default, which is
a useful feature when debugging cloudflared issues without manual intervention from users.
Users can still opt-out by disabling the feature flag.
Propagates the logger context into further locations to help provide more context for certain errors. For instance, upstream and downstream copying errors will properly have the assigned flow id attached and destination address.
## Summary
To prevent bad eyeballs and severs to be able to exhaust the quic
control flows we are adding the possibility of having a timeout
for a write operation to be acknowledged. This will prevent hanging
connections from exhausting the quic control flows, creating a DDoS.
Mads Jon Nielsen
Devin Carr
João "Pisco" Fernandes
chungthuang
GoncaloGarcia
Steven Kreitzer
Alexandru Tocar
chungthuang
sellskin
Nikita Sivukhin
Nanashi
YueYue
K.B.Dharun Krishna
João Oliveirinha
Igor Postelnik