fix: handle curl timeout failures

[taigrr]

Summary

• add explicit curl connect and total timeouts to the uptime check
• keep the workflow writing a log entry when curl exits nonzero
• render missing log entries as missing instead of down

Fixes #8

Verification

• ruby -e 'require "yaml"; YAML.load_file(".github/workflows/ping.yaml"); puts "yaml ok"'\n- git diff --check\n- simulated status mapping for 200, 302, 500, curl exit 28, and curl exit 56