v0.2.2

Fixed

• ... parts of the documentation, nothing else.

v0.2.1

Added

• Support to set a proxy selectively for one client (Prometheus/Icinga) in the configuration file if setting a proxy via the environment is not viable or desired for some reason.

v0.2.0

Added

• Properly interpret PromQL results (of time "vector"), not just single time series. Provide reasonable default outputs (plugin output/performance data) and enable improved ways of customizations for both.
• Allow customization of the unit of measurement in performance data.
• The .deb package now contains a systemd service template, so multiple instances of vec2checkd may be run in parallel.

Changed

• Use handlebars to render templates.
• Improve error handling, display the complete error chain with all the context that has been added.
• Execute every mapping/check once before entering the periodic check loop to avoid Icinga falling back to the default host/service state when vec2checkd is down for a short time, e.g. due to a restart.
• Rename Data.metric to Data.labels which can be used in plugin output and performance data templates.

Fixed

• The proper exit value is now send to Icinga when a host object is to be updated (one of 0 for "UP" or 1 for "DOWN") as other exit values (2 / 3) are not valid in this context.

v0.1.1

Changed

• Truncate the numerical value in the plugin output to two decimal places if it has a fraction part or to an integer if there is no fractional part. This is consistent with the truncation (or rounding) that Icinga2 performs on performance data. Also the plugin output looks "cleaner" when the output of a PromQL query will always be an integer (for example when the number of running K8s pods is retrieved).
• Improve some minor bits in the README
• Bump dependencies (especially rocket to v0.4.4 as v0.4.3 has been yanked on crates.io)

v0.1.0

Initial release