#!/bin/bash

################################################
#### Container Health Prometheus Exporter   ####
#### for node_exporter textfile collector   ####
####                                        ####
#### Author: Phil Connor                    ####
#### Contact: contact@mylinux.work          ####
#### Version: 1.0.0.20260309               ####
################################################

set -o pipefail

SCRIPT_NAME=$(basename "$0")
readonly SCRIPT_NAME

# Default configuration
readonly DEFAULT_NODE_DIR="/var/lib/node_exporter"
readonly DEFAULT_COLLECTION_INTERVAL=60

# Configuration variables (can be overridden by environment)
NODE_DIR=${NODE_DIR:-$DEFAULT_NODE_DIR}
COLLECTION_INTERVAL=${COLLECTION_INTERVAL:-$DEFAULT_COLLECTION_INTERVAL}
DEBUG=${DEBUG:-}

# Runtime flags
RUN_MODE="once"

handle_error() {
    local exit_code=$1
    local line_number=$2
    echo "Error: $SCRIPT_NAME failed at line $line_number with exit code $exit_code" >&2
    exit "$exit_code"
}

trap 'handle_error $? $LINENO' ERR

debug_echo() {
    if [[ -n "$DEBUG" ]]; then
        echo "[DEBUG] $*" >&2
    fi
}

show_help() {
    cat << EOF
Usage: $SCRIPT_NAME [OPTIONS]

Container health metrics collector for Prometheus node_exporter textfile directory.

Collects per-container health check status, image age, restart counts, exit codes,
and running state via docker inspect and writes them as Prometheus metrics.

OPTIONS:
    --once          Run collection once and exit (default)
    --daemon        Run continuously at COLLECTION_INTERVAL
    --help, -h      Show this help message

ENVIRONMENT VARIABLES:
    NODE_DIR                Node exporter textfile directory (default: $DEFAULT_NODE_DIR)
    COLLECTION_INTERVAL     Seconds between collections in daemon mode (default: $DEFAULT_COLLECTION_INTERVAL)
    DEBUG                   Enable debug output

EXAMPLES:
    $SCRIPT_NAME --once
    $SCRIPT_NAME --daemon
    COLLECTION_INTERVAL=30 $SCRIPT_NAME --daemon

OUTPUT:
    Writes metrics to \$NODE_DIR/textfile_collector/container_health.prom

EOF
    exit 0
}

# Parse arguments
while [[ $# -gt 0 ]]; do
    case "$1" in
        --once)     RUN_MODE="once"; shift ;;
        --daemon)   RUN_MODE="daemon"; shift ;;
        --help|-h)  show_help ;;
        *) echo "Unknown option: $1" >&2; show_help ;;
    esac
done

# Validate configuration
validate_config() {
    if ! command -v docker &>/dev/null; then
        echo "Error: docker is not installed or not in PATH" >&2
        exit 1
    fi

    local textfile_dir="${NODE_DIR}/textfile_collector"
    if [[ ! -d "$textfile_dir" ]]; then
        echo "Error: Textfile collector directory not found: $textfile_dir" >&2
        echo "Create it: sudo mkdir -p $textfile_dir" >&2
        exit 1
    fi
}

# Collect metrics for all containers
collect_all() {
    local output_dir="${NODE_DIR}/textfile_collector"
    local output_file="${output_dir}/container_health.prom"
    local temp_file
    temp_file=$(mktemp "${output_file}.XXXXXX")

    local start_time
    start_time=$(date +%s%N)
    local success=1

    debug_echo "Starting collection..."

    {
        local containers
        containers=$(docker ps -a --format '{{.Names}}')

        if [[ -z "$containers" ]]; then
            debug_echo "No containers found"
        fi

        # Per-container metrics headers
        echo "# HELP container_health_status Health check status of the container (1 for current status)."
        echo "# TYPE container_health_status gauge"
        echo "# HELP container_image_age_seconds Age of the container image in seconds."
        echo "# TYPE container_image_age_seconds gauge"
        echo "# HELP container_restart_count Number of container restarts."
        echo "# TYPE container_restart_count gauge"
        echo "# HELP container_exit_code Exit code of the container."
        echo "# TYPE container_exit_code gauge"
        echo "# HELP container_running Whether the container is running (1=running, 0=stopped)."
        echo "# TYPE container_running gauge"

        local now
        now=$(date +%s)

        while IFS= read -r container_name; do
            [[ -z "$container_name" ]] && continue

            debug_echo "Inspecting container: $container_name"

            # Extract all fields in a single docker inspect call
            local inspect_data
            inspect_data=$(docker inspect --format \
                '{{.Config.Image}}|{{if .State.Health}}{{.State.Health.Status}}{{else}}none{{end}}|{{.Created}}|{{.RestartCount}}|{{.State.ExitCode}}|{{.State.Running}}' \
                "$container_name" 2>/dev/null) || {
                debug_echo "Failed to inspect container: $container_name"
                success=0
                continue
            }

            local image health_status created restart_count exit_code running_raw
            IFS='|' read -r image health_status created restart_count exit_code running_raw <<< "$inspect_data"

            # Calculate image age in seconds
            local created_epoch image_age
            created_epoch=$(date -d "$created" +%s 2>/dev/null) || created_epoch=0
            image_age=$((now - created_epoch))

            # Convert running boolean to 0/1
            local running=0
            if [[ "$running_raw" == "true" ]]; then
                running=1
            fi

            # Health status — emit a 1 for the current status, 0 for others
            for status in healthy unhealthy starting none; do
                if [[ "$health_status" == "$status" ]]; then
                    echo "container_health_status{name=\"${container_name}\",image=\"${image}\",status=\"${status}\"} 1"
                else
                    echo "container_health_status{name=\"${container_name}\",image=\"${image}\",status=\"${status}\"} 0"
                fi
            done

            echo "container_image_age_seconds{name=\"${container_name}\",image=\"${image}\"} ${image_age}"
            echo "container_restart_count{name=\"${container_name}\",image=\"${image}\"} ${restart_count}"
            echo "container_exit_code{name=\"${container_name}\",image=\"${image}\"} ${exit_code}"
            echo "container_running{name=\"${container_name}\",image=\"${image}\"} ${running}"

        done <<< "$containers"

        # Exporter metadata
        local end_time duration
        end_time=$(date +%s%N)
        duration=$(awk "BEGIN {printf \"%.4f\", ($end_time - $start_time) / 1000000000}")

        echo ""
        echo "# HELP container_health_exporter_duration_seconds Time taken to collect metrics."
        echo "# TYPE container_health_exporter_duration_seconds gauge"
        echo "container_health_exporter_duration_seconds ${duration}"
        echo ""
        echo "# HELP container_health_exporter_last_run_timestamp Unix timestamp of last collection."
        echo "# TYPE container_health_exporter_last_run_timestamp gauge"
        echo "container_health_exporter_last_run_timestamp $(date +%s)"
        echo ""
        echo "# HELP container_health_exporter_success Whether the last collection succeeded (1=success, 0=failure)."
        echo "# TYPE container_health_exporter_success gauge"
        echo "container_health_exporter_success ${success}"

    } > "$temp_file" 2>/dev/null

    mv "$temp_file" "$output_file"

    debug_echo "Collection complete. Wrote to $output_file"
}

# Main
main() {
    validate_config

    case "$RUN_MODE" in
        once)
            collect_all
            ;;
        daemon)
            echo "$SCRIPT_NAME running in daemon mode (interval: ${COLLECTION_INTERVAL}s)"
            while true; do
                collect_all
                sleep "$COLLECTION_INTERVAL"
            done
            ;;
    esac
}

main