Configure Varnish cache invalidation with automated purging strategies for high-performance web acceleration

Install Varnish and required tools

Start by installing Varnish and tools needed for cache management and monitoring.

sudo apt update
sudo apt install -y varnish varnish-dev curl jq htop
sudo systemctl enable varnish

sudo dnf update -y
sudo dnf install -y varnish varnish-devel curl jq htop
sudo systemctl enable varnish

Configure Varnish service parameters

Set up Varnish with appropriate memory allocation and listening ports for cache invalidation operations.

[Service]
ExecStart=
ExecStart=/usr/sbin/varnishd \
  -a :80 \
  -a :6081,PROXY \
  -T localhost:6082 \
  -f /etc/varnish/default.vcl \
  -S /etc/varnish/secret \
  -s malloc,2g \
  -p feature=+http2

Create the override directory if it doesn't exist:

sudo mkdir -p /etc/systemd/system/varnish.service.d
sudo systemctl daemon-reload

Create advanced VCL configuration with cache invalidation

Configure comprehensive VCL rules supporting PURGE, BAN operations, and cache tagging for selective invalidation.

vcl 4.1;

import std;
import directors;

# Define backend servers
backend web1 {
    .host = "192.168.1.10";
    .port = "8080";
    .probe = {
        .url = "/health";
        .timeout = 2s;
        .interval = 5s;
        .window = 3;
        .threshold = 2;
    };
}

backend web2 {
    .host = "192.168.1.11";
    .port = "8080";
    .probe = {
        .url = "/health";
        .timeout = 2s;
        .interval = 5s;
        .window = 3;
        .threshold = 2;
    };
}

# Access Control List for cache management
acl purge_acl {
    "localhost";
    "127.0.0.1";
    "192.168.1.0"/24;
    "10.0.0.0"/8;
}

sub vcl_init {
    # Round-robin director for load balancing
    new web_director = directors.round_robin();
    web_director.add_backend(web1);
    web_director.add_backend(web2);
}

sub vcl_recv {
    # Set backend director
    set req.backend_hint = web_director.backend();
    
    # Handle PURGE requests
    if (req.method == "PURGE") {
        if (client.ip !~ purge_acl) {
            return (synth(405, "PURGE not allowed from " + client.ip));
        }
        
        # PURGE specific URL
        ban("req.url ~ " + req.url);
        return (synth(200, "Purged: " + req.url));
    }
    
    # Handle BAN requests for bulk invalidation
    if (req.method == "BAN") {
        if (client.ip !~ purge_acl) {
            return (synth(405, "BAN not allowed from " + client.ip));
        }
        
        if (req.http.X-Ban-Expression) {
            ban(req.http.X-Ban-Expression);
            return (synth(200, "Banned: " + req.http.X-Ban-Expression));
        } else {
            return (synth(400, "X-Ban-Expression header required"));
        }
    }
    
    # Handle tagged cache invalidation
    if (req.method == "PURGETAG") {
        if (client.ip !~ purge_acl) {
            return (synth(405, "PURGETAG not allowed from " + client.ip));
        }
        
        if (req.http.X-Cache-Tags) {
            ban("obj.http.X-Cache-Tags ~ " + req.http.X-Cache-Tags);
            return (synth(200, "Purged tags: " + req.http.X-Cache-Tags));
        } else {
            return (synth(400, "X-Cache-Tags header required"));
        }
    }
    
    # Normalize request for better caching
    if (req.url ~ "\?$") {
        set req.url = regsub(req.url, "\?$", "");
    }
    
    # Remove tracking parameters
    if (req.url ~ "(\?|&)(utm_[a-z]+|gclid|fbclid|_ga|_gid)=") {
        set req.url = regsuball(req.url, "(\?|&)(utm_[a-z]+|gclid|fbclid|_ga|_gid)=[^&]*", "");
        set req.url = regsuball(req.url, "&+", "&");
        set req.url = regsuball(req.url, "\?&", "?");
        set req.url = regsuball(req.url, "\?$", "");
    }
    
    # Handle cache bypass for specific scenarios
    if (req.http.Cache-Control ~ "no-cache" || req.http.Pragma ~ "no-cache") {
        if (client.ip ~ purge_acl) {
            set req.hash_always_miss = true;
        }
    }
    
    # Set cache key components
    set req.http.X-Forwarded-For = client.ip;
    
    return (pass) if (req.method != "GET" && req.method != "HEAD");
    return (pass) if (req.http.Authorization);
    
    return (hash);
}

sub vcl_backend_response {
    # Set cache tags from backend response
    if (beresp.http.X-Cache-Tags) {
        set beresp.http.X-Cache-Tags = beresp.http.X-Cache-Tags;
    }
    
    # Configure TTL based on content type
    if (beresp.http.Content-Type ~ "text/html") {
        set beresp.ttl = 300s;  # 5 minutes for HTML
        set beresp.grace = 1h;
    } elsif (beresp.http.Content-Type ~ "(css|js)") {
        set beresp.ttl = 1h;    # 1 hour for CSS/JS
        set beresp.grace = 6h;
    } elsif (beresp.http.Content-Type ~ "image/") {
        set beresp.ttl = 24h;   # 24 hours for images
        set beresp.grace = 48h;
    }
    
    # Handle backend errors gracefully
    if (beresp.status >= 500 && beresp.status <= 599) {
        set beresp.ttl = 0s;
        set beresp.grace = 10s;
        return (abandon);
    }
    
    # Enable streaming for large objects
    if (std.integer(beresp.http.Content-Length, 0) > 10485760) {
        set beresp.do_stream = true;
        set beresp.ttl = 3600s;
    }
    
    return (deliver);
}

sub vcl_deliver {
    # Add cache status headers
    if (obj.hits > 0) {
        set resp.http.X-Cache = "HIT";
        set resp.http.X-Cache-Hits = obj.hits;
    } else {
        set resp.http.X-Cache = "MISS";
    }
    
    # Add cache age information
    set resp.http.X-Cache-Age = obj.age;
    
    # Remove internal headers from client response
    unset resp.http.X-Cache-Tags;
    unset resp.http.Via;
    unset resp.http.X-Varnish;
    
    return (deliver);
}

sub vcl_hit {
    # Handle cache invalidation during hit
    if (req.method == "PURGE") {
        return (synth(200, "Purged"));
    }
    
    return (deliver);
}

sub vcl_miss {
    # Handle cache invalidation during miss
    if (req.method == "PURGE") {
        return (synth(404, "Not in cache"));
    }
    
    return (fetch);
}

Create cache invalidation automation scripts

Set up scripts for automated cache purging based on content changes and scheduled maintenance.

#!/bin/bash

# Varnish Cache Invalidation Script
# Usage: varnish-purge.sh [purge|ban|tag] [target]

VARNISH_HOST="localhost:80"
VARNISH_MGMT="localhost:6082"
LOG_FILE="/var/log/varnish/purge.log"

# Ensure log directory exists
sudo mkdir -p /var/log/varnish

# Logging function
log_action() {
    echo "$(date '+%Y-%m-%d %H:%M:%S') - $1" | sudo tee -a "$LOG_FILE"
}

# PURGE specific URL
purge_url() {
    local url="$1"
    if [[ -z "$url" ]]; then
        echo "Usage: $0 purge /path/to/resource"
        exit 1
    fi
    
    response=$(curl -s -X PURGE -w "%{http_code}" "http://$VARNISH_HOST$url")
    http_code=${response: -3}
    
    if [[ "$http_code" == "200" ]]; then
        log_action "PURGE SUCCESS: $url (HTTP $http_code)"
        echo "Successfully purged: $url"
    else
        log_action "PURGE FAILED: $url (HTTP $http_code)"
        echo "Failed to purge: $url (HTTP $http_code)"
        exit 1
    fi
}

# BAN with expression
ban_expression() {
    local expression="$1"
    if [[ -z "$expression" ]]; then
        echo "Usage: $0 ban 'req.url ~ /api/'"
        exit 1
    fi
    
    response=$(curl -s -X BAN -H "X-Ban-Expression: $expression" -w "%{http_code}" "http://$VARNISH_HOST/")
    http_code=${response: -3}
    
    if [[ "$http_code" == "200" ]]; then
        log_action "BAN SUCCESS: $expression (HTTP $http_code)"
        echo "Successfully banned: $expression"
    else
        log_action "BAN FAILED: $expression (HTTP $http_code)"
        echo "Failed to ban: $expression (HTTP $http_code)"
        exit 1
    fi
}

# PURGE by cache tags
purge_tags() {
    local tags="$1"
    if [[ -z "$tags" ]]; then
        echo "Usage: $0 tag 'user:123,content:article'"
        exit 1
    fi
    
    response=$(curl -s -X PURGETAG -H "X-Cache-Tags: $tags" -w "%{http_code}" "http://$VARNISH_HOST/")
    http_code=${response: -3}
    
    if [[ "$http_code" == "200" ]]; then
        log_action "PURGETAG SUCCESS: $tags (HTTP $http_code)"
        echo "Successfully purged tags: $tags"
    else
        log_action "PURGETAG FAILED: $tags (HTTP $http_code)"
        echo "Failed to purge tags: $tags (HTTP $http_code)"
        exit 1
    fi
}

# Main script logic
case "$1" in
    purge)
        purge_url "$2"
        ;;
    ban)
        ban_expression "$2"
        ;;
    tag)
        purge_tags "$2"
        ;;
    *)
        echo "Usage: $0 {purge|ban|tag} [target]"
        echo "Examples:"
        echo "  $0 purge /api/users/123"
        echo "  $0 ban 'req.url ~ /old-content/'"
        echo "  $0 tag 'user:123,content:article'"
        exit 1
        ;;
esac

Make the script executable:

sudo chmod 755 /usr/local/bin/varnish-purge.sh

Create cache monitoring and metrics script

Set up automated monitoring to track cache hit rates, invalidation frequency, and performance metrics.

#!/bin/bash

# Varnish Cache Monitoring Script
# Collects and reports cache performance metrics

VARNISH_MGMT="localhost:6082"
METRICS_LOG="/var/log/varnish/metrics.log"
ALERT_THRESHOLD=80  # Alert if cache hit rate drops below 80%

# Get cache statistics
get_cache_stats() {
    local stats=$(varnishstat -1 -f MAIN.cache_hit -f MAIN.cache_miss -f MAIN.n_object -f MAIN.n_expired -f MAIN.n_lru_nuked)
    
    cache_hits=$(echo "$stats" | grep 'MAIN.cache_hit' | awk '{print $2}')
    cache_misses=$(echo "$stats" | grep 'MAIN.cache_miss' | awk '{print $2}')
    objects_cached=$(echo "$stats" | grep 'MAIN.n_object' | awk '{print $2}')
    objects_expired=$(echo "$stats" | grep 'MAIN.n_expired' | awk '{print $2}')
    objects_lru_nuked=$(echo "$stats" | grep 'MAIN.n_lru_nuked' | awk '{print $2}')
    
    # Calculate hit rate
    total_requests=$((cache_hits + cache_misses))
    if [[ $total_requests -gt 0 ]]; then
        hit_rate=$((cache_hits * 100 / total_requests))
    else
        hit_rate=0
    fi
    
    # Log metrics
    timestamp=$(date '+%Y-%m-%d %H:%M:%S')
    metrics_line="$timestamp,hits:$cache_hits,misses:$cache_misses,hit_rate:$hit_rate%,objects:$objects_cached,expired:$objects_expired,lru_nuked:$objects_lru_nuked"
    
    echo "$metrics_line" | sudo tee -a "$METRICS_LOG"
    
    # Alert on low hit rate
    if [[ $hit_rate -lt $ALERT_THRESHOLD ]]; then
        echo "WARNING: Cache hit rate dropped to $hit_rate% (threshold: $ALERT_THRESHOLD%)"
        logger -t varnish-monitor "Cache hit rate alert: $hit_rate%"
    fi
    
    # Display current stats
    echo "=== Varnish Cache Statistics ==="
    echo "Cache Hits: $cache_hits"
    echo "Cache Misses: $cache_misses"
    echo "Hit Rate: $hit_rate%"
    echo "Objects Cached: $objects_cached"
    echo "Objects Expired: $objects_expired"
    echo "Objects LRU Nuked: $objects_lru_nuked"
}

# Get ban list information
get_ban_stats() {
    echo "\n=== Active Bans ==="
    varnishadm -T "$VARNISH_MGMT" ban.list | head -10
}

# Display top URLs by cache status
get_top_urls() {
    echo "\n=== Top Cached URLs (last 1000 requests) ==="
    varnishlog -n default -d -i ReqURL,VCL_return -m RxURL:. | \
    head -1000 | \
    grep -E 'ReqURL|VCL_return' | \
    paste - - | \
    awk '{print $4, $8}' | \
    sort | uniq -c | sort -rn | head -20
}

# Main execution
echo "Varnish Cache Monitor - $(date)"
get_cache_stats
get_ban_stats
get_top_urls

Make the monitoring script executable:

sudo chmod 755 /usr/local/bin/varnish-monitor.sh

Set up automated cache warming

Create a cache warming system that preloads frequently accessed content after cache invalidations.

#!/bin/bash

# Varnish Cache Warming Script
# Preloads critical URLs after cache invalidation

WARM_URLS_FILE="/etc/varnish/warm-urls.txt"
VARNISH_HOST="localhost:80"
CONCURRENCY=10
TIMEOUT=30

# Create URLs file if it doesn't exist
if [[ ! -f "$WARM_URLS_FILE" ]]; then
    sudo tee "$WARM_URLS_FILE" << EOF
# Critical URLs to warm after cache invalidation
# One URL per line
/
/api/status
/popular-content
/category/technology
/category/news
EOF
fi

# Warm cache function
warm_cache() {
    local url="$1"
    local start_time=$(date +%s.%N)
    
    response=$(curl -s -w "%{http_code}:%{time_total}" -m "$TIMEOUT" "http://$VARNISH_HOST$url")
    http_code=$(echo "$response" | cut -d':' -f1)
    response_time=$(echo "$response" | cut -d':' -f2)
    
    local end_time=$(date +%s.%N)
    local duration=$(echo "$end_time - $start_time" | bc)
    
    if [[ "$http_code" == "200" ]]; then
        echo "WARM SUCCESS: $url (${response_time}s)"
    else
        echo "WARM FAILED: $url (HTTP $http_code)"
    fi
}

# Export function for parallel execution
export -f warm_cache
export VARNISH_HOST TIMEOUT

# Main warming process
echo "Starting cache warming at $(date)"
echo "Using $CONCURRENCY concurrent connections"
echo "Timeout: ${TIMEOUT}s per request"
echo ""

# Process URLs in parallel
if [[ -f "$WARM_URLS_FILE" ]]; then
    grep -v '^#' "$WARM_URLS_FILE" | grep -v '^$' | \
    xargs -n 1 -P "$CONCURRENCY" -I {} bash -c 'warm_cache "$@"' _ {}
else
    echo "Warning: URLs file not found: $WARM_URLS_FILE"
    exit 1
fi

echo ""
echo "Cache warming completed at $(date)"

Make the warming script executable:

sudo chmod 755 /usr/local/bin/varnish-warm.sh

Configure systemd timers for automated monitoring

Set up systemd services and timers for regular cache monitoring and maintenance.

[Unit]
Description=Varnish Cache Monitoring
After=varnish.service

[Service]
Type=oneshot
ExecStart=/usr/local/bin/varnish-monitor.sh
User=varnish
Group=varnish

[Unit]
Description=Run Varnish Cache Monitoring every 5 minutes
Requires=varnish-monitor.service

[Timer]
OnCalendar=*:0/5
Persistent=true

[Install]
WantedBy=timers.target

Enable and start the monitoring timer:

sudo systemctl daemon-reload
sudo systemctl enable varnish-monitor.timer
sudo systemctl start varnish-monitor.timer

Start and verify Varnish configuration

Reload the Varnish configuration and ensure all invalidation features are working correctly.

sudo systemctl restart varnish
sudo systemctl status varnish
varnishd -C -f /etc/varnish/default.vcl

Test PURGE functionality

# Fetch a page to cache it
curl -I http://localhost/

# PURGE the specific URL
curl -X PURGE http://localhost/

# Use the purge script
/usr/local/bin/varnish-purge.sh purge /api/users

Test BAN operations

# BAN all URLs matching a pattern
curl -X BAN -H "X-Ban-Expression: req.url ~ /api/" http://localhost/

# Use the purge script for BAN
/usr/local/bin/varnish-purge.sh ban 'req.url ~ /old-content/'

Test cache tag invalidation

# PURGE by cache tags
curl -X PURGETAG -H "X-Cache-Tags: user:123" http://localhost/

# Use the purge script for tags
/usr/local/bin/varnish-purge.sh tag 'content:article,category:tech'

Monitor cache performance

# Run monitoring script
/usr/local/bin/varnish-monitor.sh

# Check Varnish statistics
varnishstat

# View active bans
varnishadm ban.list

# Check cache logs
sudo tail -f /var/log/varnish/purge.log

Implement webhook-based invalidation

Create a webhook endpoint that triggers cache invalidation when content management systems update content.

#!/usr/bin/env python3

import json
import subprocess
import sys
from http.server import HTTPServer, BaseHTTPRequestHandler
from urll