832334655c
The Docker health check was attempting to bind to ports already in use by the running application, causing "address already in use" errors. This fix adds a health check mode that skips network binding operations while still validating output configuration (listen addresses, paths, etc.). Changes: - Add health check mode flag in pkg/poller/outputs.go - Update prometheus and webserver DebugOutput() to skip port binding in health check mode - Maintain full configuration validation without network conflicts Fixes #892 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com> |
||
|---|---|---|
| .. | ||
| LICENSE | ||
| README.md | ||
| commands.go | ||
| config.go | ||
| inputs.go | ||
| logger.go | ||
| outputs.go | ||
| start.go | ||
| testutil.go | ||
README.md
poller
UniFi Poller Core
This module ties the inputs together with the outputs.
Aggregates metrics on request. Provides CLI app and args parsing.
Ideal
This library has no notion of "UniFi" or controllers, or Influx, or Prometheus.
This library simply provides an input interface and an output interface.
Each interface uses an []any type, so any type of data can be used.
That is to say, you could write input and output plugins that work with, say,
Cisco gear, or any other network (or even non-network) data. The existing plugins
should provide ample example of how to use this library, but at some point the
godoc will improve.
Features
- Automatically unmarshal's plugin config structs from config file and/or env variables.
- Initializes all "imported" plugins on startup.
- Provides input plugins a Logger, requires an interface for Metrics and Events retrieval.
- Provides Output plugins an interface to retrieve Metrics and Events, and a Logger.
- Provides automatic aggregation of Metrics and Events from multiple sources.