criblio
Timeouts seen with appscope/edge integration
What was reported: started edge inside a docker container started top process on host attached to process after some amount of time (less than a minute), the inspect stopped showing the process as scoped (with no changes made) history doesn’t show a duration, which may imply that it’s still kind of running.
the scope attach command reported Inspect PID fails: timeout with consume the message
The machine where it was seen: "11th Gen Intel Core i7-1165G7 @ 2.80GHz x 8"
We thought this was resolved by https://github.com/criblio/appscope/pull/1579, but appears to persist.
Today I found pretty easy reproduction steps:
# Try to scope process where event destination is not available (nobody listens)
scope run -e tcp://localhost:9999 -- nc -lp 10001 &
# right after it try to run
scope inspect <nc_pid>
<Inspect failure: timeout with consume the message>
# the scope inspect will work correctly after 4~5 seconds
I susspect the reporting thread is busy with retry/reconnection try that it is not able to hit IPC handling code
