ci: fix flaky workflows #438
Description
A number of jobs fail fairly regularly with messages like the logs attached below. In particular, test-coverage and integration-tests. Usually only a subset of the integration tests fail. Rerunning the jobs sometimes lets them succeed. Example: https://github.com/posit-dev/connectapi/actions/runs/16762656430/job/47461346309?pr=437
connect Pulled
Network ci_default Creating
Network ci_default Created
Container ci-connect-2 Creating
Container ci-connect-1 Creating
Container ci-connect-1 Created
Container ci-connect-2 Created
Container ci-connect-2 Starting
Container ci-connect-2 Started
Container ci-connect-1 Starting
Container ci-connect-1 Started
docker compose: started!
docker: getting list of containers...
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
cc8f4f5a2421 ghcr.io/rstudio/rstudio-connect:jammy-2024.03.0 "tini -- /usr/local/…" Less than a second ago Up Less than a second 0.0.0.0:32768->3939/tcp, [::]:32768->3939/tcp ci-connect-2
e2df01a45872 ghcr.io/rstudio/rstudio-connect:jammy-2024.03.0 "tini -- /usr/local/…" Less than a second ago Up Less than a second 0.0.0.0:32769->3939/tcp, [::]:32769->3939/tcp ci-connect-1
docker: got ports 32768 and 32769
connect: sleeping - waiting for connect to start
connect: creating first admin...
Error in curl::curl_fetch_memory(url, handle = handle) :
Couldn't connect to server [localhost]:
Failed to connect to localhost port 32768 after 0 ms: Couldn't connect to server
Calls: <Anonymous> ... request_fetch.write_memory -> <Anonymous> -> raise_libcurl_error
In addition: Warning message:
DO NOT USE IN PRODUCTION - The build_test_env function is for internal testing purposes only
This warning is displayed once per session.
Execution halted