Return empty list instead of 0 in error case
Fix Prometheus metrics output formatting
Allow a certain threshold of errors before reporting invalid cert
A `max_error_count` config entry can be provided, if fewer than that
amount of errors occur, the certificate validity will be reported as the
"old" value (before the first error. If more than `max_error_count`
errors occur in a row, validity is reported as `0` (before this commit,
that was the default behaviour after the first error).
Use timeout for TLS connection
Add .plt file to .gitignore
Fix format string for config path logging
Use foreach instead of map to make dialyzer happy
Use application:set_env/3 instead of /2 for compatibility with OTP < 21
Externalize configuration
Path to the configuration file is now passed via an environment
variable. This makes it possible to run the same release with different
configurations.
Handle connection errors in scanner instead of crashing
Start scanner processes supervised
Start the individual scanner processes as part of the app's supervision
tree, so they are restarted in case of a crash. Also increase the time
before the first scan a bit.
Do not crash metrics handler, when scanner process does not exist
Add initial documentation
Configure Sourcehut CI builds