Context
I've noticed many recent GH Action failures were caused by timeout during setup-nextstrain-cli (flu, forecasts-ncov, hmpv, ncov, rsv, seasonal-cov):
Prepare all required actions
Run ./.git/nextstrain/.github/actions/setup-nextstrain-cli
Run curl -fsSL --proto '=https' https://nextstrain.org/cli/installer/"$OS" | bash
curl: (28) Failed to connect to nextstrain.org port 443 after 300791 ms: Timeout was reached
Error: Process completed with exit code 28.
Description
It'd be nice for setup-nextstrain-cli to automatically retry downloading the Nextstrain CLI if it runs into transient errors. We'd likely want the number of retries to be a input that can be set per workflow.
For linux/mac, curl has a built-in --retry flag. For windows, I'm unfamiliar with Invoke-RestMethod, but it has a -MaximumRetryCount parameter which should work here.
Context
I've noticed many recent GH Action failures were caused by timeout during
setup-nextstrain-cli(flu, forecasts-ncov, hmpv, ncov, rsv, seasonal-cov):Description
It'd be nice for
setup-nextstrain-clito automatically retry downloading the Nextstrain CLI if it runs into transient errors. We'd likely want the number of retries to be a input that can be set per workflow.For linux/mac, curl has a built-in
--retryflag. For windows, I'm unfamiliar withInvoke-RestMethod, but it has a-MaximumRetryCountparameter which should work here.