Buyer’s Guide - August 2026

The 10 best Varnish Cache monitoring tools, ranked

Q: How do I monitor Varnish with Prometheus and Grafana?

The community prometheus_varnish_exporter scrapes varnishstat -j output and exposes metrics on port 9131. Grafana Cloud’s Varnish Cache integration builds on it with 2 pre-built dashboards and 6 alerts covering hit rate, evictions, saturation, and backend health. For access logs, varnishncsa feeds the Varnish logs dashboard. One exporter instance is needed per Varnish instance, selected with the -n flag in multi-instance setups.

Q: What is varnishstat and how do I use it?

varnishstat is the CLI tool that exposes Varnish’s in-memory performance counters. Use -1 for one-shot output, -j for JSON, and -f to filter specific counters. For request-level detail, varnishlog, varnishtop, varnishhist, and varnishsizes complement the counters. Most monitoring integrations on this page parse varnishstat output under the hood rather than instrumenting Varnish directly.

Varnish monitoring lives or dies on varnishstat: cache hit ratio, hitpass and hitmiss, thread queue length, backend health, and storage pressure. This ranking grades ten tools on how deeply they expose those counters, how fast they get you to a useful dashboard, and whether the bill grows with nodes or with metric volume.

Start with Netdata (free) Jump to rankings

Why this list exists

Varnish is not a generic process. The metrics that tell you whether your cache is doing its job live inside varnishstat: cache hit ratio, hitpass and hitmiss, client requests, session drops, thread queue length, backend health, and per-storage usage. The mistake buyers make is treating Varnish like any other daemon and monitoring only CPU and memory. That catches a crashed process but misses the failures that actually hurt: a collapsing hit ratio, saturated worker threads, and backends going unhealthy while the server looks fine.

Three dimensions decide the outcome when shortlisting a Varnish monitoring tool:

Metric depth. Does the tool parse varnishstat properly, with per-backend and per-storage detail, or does it scrape a handful of counters through a plugin?
Resolution and alerting. A cache storm develops in seconds. Tools polling at 60 seconds will show you the aftermath, not the event. Pre-built alerts for low hit rate, thread saturation, and unhealthy backends matter more than another dashboard.
Pricing shape. High-frequency cache telemetry is exactly the kind of data that makes per-GB, per-series, and per-metric bills grow. Per-node pricing and open-source self-hosting behave very differently at scale.

One note on how to read this page: we do not quote competitor list prices. Pricing pages change, negotiated rates differ, and a dollar figure copied into a listicle is stale the week it publishes. Instead we describe each tool’s pricing shape, meaning which meter the bill grows on, and link the official pricing page so you can check current numbers yourself. For hands-on configuration guidance, our Varnish monitoring guides cover the operational side in detail.

The shortlist was assembled from tools with documented, working Varnish integrations: vendor-maintained collectors, official extensions, and widely used community plugins. We excluded tools with no Varnish-specific coverage and verified each integration against current vendor documentation as of August 2026.

Varnish metric depth carries the most weight because it is the whole point of the category. Time to value comes second: a tool that auto-detects Varnish beats one that requires assembling an exporter, templates, and dashboards. Resolution and alerting follow, because cache incidents are fast. Pricing predictability matters more here than in most categories, since Varnish generates dense, high-frequency telemetry.

Tester credit

Compiled by the Netdata team - Updated August 12, 2026

Sources verified

Scoring criteria

Varnish metric depth and fidelity 25%
varnishstat counters that matter: hit ratio, threads, backends, storage
Time to value 20%
Auto-detection and pre-built dashboards beat manual assembly
Data resolution 15%
Per-second and 10-second collection catches cache storms
Alerting and anomaly detection 15%
Pre-built alerts for hit rate, threads, evictions, backends
Pricing predictability 15%
Per-node beats per-GB and per-series for dense cache telemetry
Ecosystem correlation 10%
Varnish metrics alongside host, container, and app data

Vendor 01 / 10 · #netdata

Netdata

Real-time infrastructure monitoring with an open-source agent that auto-detects Varnish on hosts and in Docker.

Best for

Teams that want Varnish metrics correlated with the full host and container stack in one pane
Buyers who want predictable per-node pricing without per-GB or per-series surprises
Operators who want zero-config auto-detection of Varnish instances, including in Docker

Pricing

Per-node: Netdata Cloud Business starts at $4.50/node/month on annual plans, decreasing as node count grows
Free Cloud tier for up to 5 nodes; agents are open source (AGPL) and remain free
Unlimited metrics, logs, users, and retention - no per-GB, per-metric, or per-series charges

Pros

Auto-detects Varnish instances on hosts and in Docker containers with zero configuration
Collects varnishstat-based metrics with per-backend (VBE) and per-storage (SMA, SMF, MSE) detail
Supports both open-source Varnish-Cache and commercial Varnish-Plus
Cache hit ratio charts in both total and current-poll (delta) form
ML-based anomaly detection on every metric, with a 99% false-positive reduction claim
800+ integrations for correlating Varnish with hosts, containers, and applications

Where teams pair it

The Varnish collector defaults to a 10-second interval (configurable), not the per-second resolution used for system metrics
No alerts are configured by default for the Varnish integration - you create your own alert rules

Verdict

Netdata leads this category because it removes the two biggest sources of friction: setup and billing. Install the agent and Varnish appears, with per-backend and per-storage varnishstat detail, alongside the host, container, and application metrics you need to explain a cache incident. Per-node pricing with unlimited metrics means collecting dense Varnish telemetry never changes the bill. The honest caveats: Varnish charts poll at 10 seconds by default rather than per-second, and you will build your own alert rules since none ship with the integration. If request-level log analysis is a hard requirement, pair it with a log tool or the journal path.

Start free -> See Varnish monitoring ->

Vendor 02 / 10 · #grafana

Grafana Cloud + Prometheus

Grafana Cloud’s Varnish Cache integration combines a community Prometheus exporter with pre-built dashboards and alerts.

Best for

Teams already standardized on Prometheus and Grafana
Users who want pre-built Varnish dashboards and alerts without building from scratch
Organizations that want varnishncsa logs and metrics in one UI

Pricing

Per-active-series for metrics, per-GB for logs, per-user for visualization
Free tier with limited active series and retention; pay-as-you-go above it
Bill grows with metric cardinality, log volume, and number of users

Pros

Grafana-maintained integration with 2 pre-built dashboards (Varnish overview, Varnish logs) and 6 alerts
Alerts cover low cache hit rate, high memory usage, eviction rate, saturation, dropped sessions, and unhealthy backends
Collects metrics via the community varnish_exporter (varnishstat -j) and logs via varnishncsa
Integration changelog shows active maintenance through June 2026
Runs on Grafana Cloud or self-hosted Grafana Enterprise

Cons

Requires assembling and operating the community varnish_exporter plus Grafana Alloy
The exporter is a community project with no compatibility promises across Varnish versions
Per-active-series pricing punishes high-cardinality Varnish metrics
One exporter instance is needed per Varnish instance for multi-instance setups

Verdict

This is the strongest purpose-built Varnish package after Netdata: real dashboards, real alerts, and varnishncsa log collection, all maintained by Grafana. The friction is assembly. You run the community exporter per Varnish instance, wire it through Alloy, and accept that a community exporter makes no compatibility promises when Varnish versions change. For Prometheus-standardized teams that trade is routine. For everyone else, per-active-series metering on dense cache metrics is the quiet risk.

Read full Netdata vs Grafana -> Grafana pricing ->

Vendor 03 / 10 · #datadog

Datadog

SaaS observability platform with a maintained Agent check that parses varnishstat and varnishadm output.

Best for

Teams already on Datadog who want Varnish in the same UI as APM, logs, and RUM
Organizations that need backend health checks via varnishadm
Enterprises that want a fully managed SaaS with no self-hosted components

Pricing

Per-host for infrastructure monitoring, per-GB for logs and APM spans, per-metric for custom metrics
Free tier limited to a small number of hosts with short retention
Bill grows with host count, log volume, APM span volume, and custom metrics

Pros

Actively maintained Agent check (version 4.4.1) included in the Datadog Agent package
Very deep metric coverage: locks, memory pools, storage, backends, bans, sessions, threads, workspaces
Backend health service checks via varnishadm
Log collection from varnishncsa supported
Supports Varnish 3.x through 6.x with version-aware metric handling

Cons

Per-host plus per-GB plus per-metric pricing makes the bill hard to predict
Requires the Datadog Agent on each Varnish host and sudoers configuration for varnishadm
No pre-built Varnish dashboards in the core integration - dashboards are community-built
Metric names vary by Varnish version, complicating long-term dashboards

Verdict

Datadog’s Varnish check is the deepest vendor-maintained Agent integration on this list, and the only one with backend health service checks through varnishadm. If you already pay for Datadog, enabling it is obvious. What you do not get is packaging: dashboards and alerts are yours to build, and the three-meter pricing model (hosts, logs, custom metrics) means a chatty Varnish fleet costs more each time you turn up collection detail.

Read full Netdata vs Datadog -> Datadog pricing ->

Vendor 04 / 10 · #newrelic

New Relic

Usage-based observability platform with an on-host Varnish integration that reports varnishstat metrics and inventory.

Best for

Teams that want Varnish metrics alongside APM, browser, and infrastructure data
Organizations that prefer a free tier with generous ingest limits
Users who want guided install for the on-host integration

Pricing

Usage-based: per-GB data ingest plus per-user tiers
Free tier with a generous monthly data ingest allowance
Bill grows with data ingest volume and number of paid users

Pros

Official on-host integration (nri-varnish) with guided install via the New Relic CLI
Reports both metrics and inventory from Varnish instances
Broad metric coverage: backends, bans, cache, ESI, fetch, locks, sessions, storage, threads, workspaces
Open-source integration code on GitHub
Quickstart provides a pre-built Varnish dashboard

Cons

Usage-based pricing (per-GB ingest plus per-user) makes costs scale with data volume
Requires the New Relic infrastructure agent on each Varnish host
Quickstart dashboard has no bundled alerts
Docker container monitoring requires a wrapper workaround for varnishstat

Verdict

New Relic’s nri-varnish is a solid official integration with a genuinely easy guided install and a quickstart dashboard that gets you to visibility fast. Two gaps hold it back for Varnish-focused teams: no bundled alerts, so you write your own thresholds, and a per-GB ingest model where heavier collection costs more. Containerized Varnish also needs a wrapper to reach varnishstat, which adds operational detail the docs ask you to handle yourself.

Read full Netdata vs New Relic -> New Relic pricing ->

Vendor 05 / 10 · #checkmk

Checkmk

Open-core monitoring platform with native varnishstat-based checks for cache hit ratio, backends, workers, and storage.

Best for

IT operations teams already running Checkmk for host and network monitoring
Organizations that want open-source self-hosting with an optional commercial tier
Teams that prefer agent-based checks with threshold alerting

Pricing

Open-core: free Community edition (limited services), commercial subscriptions per monitored service
Commercial editions also meter custom metrics and synthetic tests
Bill grows with number of monitored services and custom metrics

Pros

Native Varnish checks built into the agent: uptime, cache hit ratio, backend success ratio, worker, cache, fetch, backend, objects
Checks parse varnishstat -1 output with no extra exporter to run
Open-core model with a free Community edition
Threshold-based alerting on each Varnish check

Cons

Varnish checks are polled at Checkmk’s standard interval (typically 60s), not real-time
No pre-built Varnish dashboards beyond the check graphs
Free edition is capped by service count
Varnish log analysis is out of scope

Verdict

Checkmk is the rare open-core platform with Varnish checks built into the agent itself, not bolted on through a community plugin. You get threshold alerting on hit ratio and backend health out of the box, and the Community edition is a real self-hosted option. The tradeoff is resolution and presentation: 60-second polling misses short cache storms, and visualization stops at per-check graphs. Solid for IT ops estates, thin for cache performance engineering.

Read full Netdata vs Checkmk -> Checkmk pricing ->

Vendor 06 / 10 · #zabbix

Zabbix

Open-source monitoring platform with community Varnish templates that collect varnishstat counters via the Zabbix agent.

Best for

Organizations committed to open-source self-hosted monitoring
Teams with Zabbix expertise who can assemble community templates
Budget-conscious buyers who want no per-metric or per-GB charges

Pricing

Open source, self-hosted - you run and operate it yourself
Commercial support subscriptions per node
Optional Zabbix Cloud SaaS with subscription pricing

Pros

Official Zabbix integration page lists Varnish templates for the community
Templates parse varnishstat output via Zabbix agent user parameters
Active community template for Varnish Cache Plus (allenta)
Fully open source with no license fees for self-hosting
Scales to very large fleets with Zabbix proxies

Cons

Varnish support relies on community templates, not a vendor-maintained integration
Template quality and Varnish version compatibility vary
No pre-built Varnish dashboards beyond template graphs
Requires Zabbix server, agent, and template assembly effort

Verdict

Zabbix can monitor Varnish well in the hands of a team that already runs it. The community templates parse varnishstat through agent user parameters, there are no license fees, and proxies scale to large fleets. But this is DIY monitoring: template quality varies, nobody at Zabbix maintains the Varnish integration, and you assemble graphs and triggers yourself. The license savings are real, and so is the operational cost that replaces them.

Read full Netdata vs Zabbix -> Zabbix pricing ->

Vendor 07 / 10 · #dynatrace

Dynatrace

Enterprise observability platform with a Varnish Cache extension that executes varnishstat and feeds metrics into Dynatrace.

Best for

Large enterprises already standardized on Dynatrace
Teams that want Varnish metrics correlated with full-stack distributed tracing
Organizations with budget for a premium consumption-priced platform

Pricing

Consumption-based: per-host (sized by memory), per-GiB logs, per-metric-datapoint, per-RUM-session meters
No free tier; public playground sandbox only
Bill grows with host memory size, log ingest, metric datapoints, and RUM or synthetic volumes

Pros

Official Varnish Cache extension documented in Dynatrace Docs
Extension executes varnishstat and sends metrics to Dynatrace
OneAgent provides host-level CPU, memory, and network context alongside Varnish metrics
Vendor-published tutorials for Varnish and Dynatrace integration

Cons

Consumption pricing across multiple meters is the least predictable model for Varnish fleets
Varnish metrics require the extension plus varnishstat path configuration
No free tier for evaluation
Metric push requires custom scripting for some setups (varnishstat to the Dynatrace API)

Verdict

Dynatrace covers Varnish through an official extension, and inside an existing Dynatrace estate the correlation with full-stack tracing is genuinely valuable. Outside that context the case is weak: the extension needs path configuration, some setups fall back to custom scripting against the API, there is no free tier to evaluate with, and consumption metering on high-frequency cache metrics is the hardest pricing shape in this list to forecast.

Read full Netdata vs Dynatrace -> Dynatrace pricing ->

Vendor 08 / 10 · #site24x7

Site24x7

SaaS monitoring platform with a Varnish Cache plugin reporting cache hits, misses, worker threads, and dropped sessions.

Best for

SMBs and web teams that want SaaS monitoring without self-hosting
Teams already using Site24x7 for website uptime and server monitoring
Users who want a simple plugin install on the Site24x7 Linux agent

Pricing

Per-monitor and per-device licensing with plan tiers
No permanent free tier; 30-day trial
Bill grows with number of monitors, hosts, and add-ons (logs, RUM, synthetics)

Pros

Official Varnish Cache plugin with documented metrics: cache_hit, cache_miss, n_wrk_create, n_wrk_queued, sess_pipe_overflow
Plugin runs on the Site24x7 Linux agent with automatic execution within five minutes
Plugin code is customizable and open on GitHub
Part of a broader website, server, and application monitoring platform

Cons

Plugin covers a narrow metric set compared to full varnishstat parsing
Requires the Python psycopg2 module as a prerequisite
Per-monitor licensing means Varnish metrics consume monitor licenses
No pre-built Varnish dashboards beyond the plugin charts

Verdict

Site24x7’s Varnish plugin is the simplest install on this list: drop it on the Linux agent and core cache health metrics appear within minutes. That simplicity is also the ceiling. The metric set is a narrow slice of varnishstat, there is no hit ratio decomposition or per-backend detail, and every Varnish plugin instance spends monitor licenses. Reasonable for an SMB watching one cache; insufficient for performance-tuning a fleet.

Read full Netdata vs Site24x7 -> Site24x7 pricing ->

Vendor 09 / 10 · #logicmonitor

LogicMonitor

SaaS IT operations platform with out-of-the-box Varnish Cache LogicModules for performance dashboards.

Best for

IT operations teams that want Varnish monitored alongside network and infrastructure devices
Organizations using LogicMonitor’s LogicModule ecosystem
Teams that prefer vendor-maintained monitoring modules

Pricing

Per-device (hybrid units) with tiered platform packages
No permanent free tier; 15-day trial
Bill grows with number of monitored devices, cloud resources, and pods

Pros

Official Varnish Cache integration with LogicModules out of the box
Monitors client requests, connections, cache hits and misses, hitpass, and backend health
Dashboards built from LogicModule data for IT operations
Part of a broad infrastructure monitoring platform

Cons

Varnish coverage is a LogicModule set, not a dedicated product with Varnish-specific dashboards
Per-device hybrid-unit pricing makes a fleet of Varnish hosts a significant line item
No Varnish log collection
Metric depth is narrower than full varnishstat parsing

Verdict

LogicMonitor’s Varnish LogicModules are vendor-maintained and cover the operational basics: requests, connections, hits and misses, hitpass, backend health. For an IT ops team monitoring Varnish next to switches and servers, that is adequate. The limits show when you tune cache performance: the metric set is basic, there is no log collection, and per-device pricing means every Varnish host is a line item whether it is a cache tier or a database.

Read full Netdata vs LogicMonitor -> LogicMonitor pricing ->

Vendor 10 / 10 · #icinga

Icinga

Open-source monitoring platform using the check_varnish plugin to alert on varnishstat counters and cache hit ratio.

Best for

Open-source shops already running Icinga 2
Teams that want threshold alerting on varnishstat counters
Organizations that prefer Nagios-style plugins

Pricing

Open source, self-hosted - you run and operate it yourself
Commercial support and managed services available from Icinga GmbH
No per-host or per-metric license fees

Pros

check_varnish plugin from the official varnish-nagios repository
Icinga plugin directory lists Varnish monitoring plugins
Monitors cache hit ratio, backend health, and varnishstat fields with warning and critical thresholds
Fully open source with no license fees
Community examples show multi-instance Varnish monitoring with Grafana graphing

Cons

Varnish support depends on community plugins, not a vendor-maintained integration
No built-in Varnish dashboards - graphing requires Grafana or another TSDB
Plugin compatibility issues with Varnish 6.5+ output changes require community fixes
Polling at Icinga check interval, typically 60s or longer

Verdict

Icinga plus check_varnish is the classic open-source alerting path for Varnish: thresholds on hit ratio and backend health, no license fees, full control. It is also the most manual option here. The plugin has known compatibility friction with newer Varnish output, dashboards mean wiring in Grafana yourself, and 60-second checks are alerting on symptoms rather than watching the cache behave. Choose it when Icinga is already your standard, not as a fresh Varnish monitoring decision.

Read full Netdata vs Icinga -> Icinga pricing ->

Frequently asked questions

What is the best way to monitor Varnish Cache?

The best way to monitor Varnish Cache is with a tool that parses varnishstat counters directly, such as Netdata, Datadog, New Relic, Checkmk, or Zabbix. varnishstat is the source of truth for Varnish performance counters, and nearly every serious monitoring tool parses it under the hood. Prometheus users typically run a varnish exporter that scrapes varnishstat -j output. When choosing, look for auto-detection of instances, cache hit ratio visibility, backend health monitoring, and pre-built alerting rather than raw counter collection alone.

What Varnish metrics should I monitor?

Start with the cache hit ratio, decomposed into hit, miss, hitpass, and hitmiss so you can see why requests bypass the cache. Track client requests and session connections and drops. Watch thread queue length and thread creation failures, which signal saturation before latency does. Monitor backend health: unhealthy, busy, and failed connections. Finally, track expired objects and LRU nuked objects as cache pressure indicators, plus storage usage per storage engine.

How do I monitor Varnish with Prometheus and Grafana?

The community prometheus_varnish_exporter scrapes varnishstat -j output and exposes metrics on port 9131. Grafana Cloud’s Varnish Cache integration builds on it with 2 pre-built dashboards and 6 alerts covering hit rate, evictions, saturation, and backend health. For access logs, varnishncsa feeds the Varnish logs dashboard. One exporter instance is needed per Varnish instance, selected with the -n flag in multi-instance setups.

What is a good Varnish cache hit ratio?

95% or higher is commonly cited as achievable for genuinely cacheable content. But the decomposition matters as much as the headline number: hitpass and hitmiss tell you which requests are deliberately or accidentally bypassing the cache. The right ratio depends on your workload and VCL configuration, not a universal target. Monitor the ratio as a trend over time and alert on meaningful drops rather than chasing a single number.

How do I set up Varnish alerting?

To set up Varnish alerting, configure alerts for low cache hit rate, high eviction rate, thread saturation, sessions dropping, and unhealthy backends. Grafana Cloud ships prebuilt Varnish alerts out of the box with its integration. Datadog provides backend health service checks via varnishadm. Netdata requires you to create your own Varnish alert rules, since none are configured by default for the integration.

Does Netdata monitor Varnish Cache?

Yes, via the go.d varnish collector. It auto-detects Varnish instances on hosts and in Docker containers with zero configuration, and supports both open-source Varnish-Cache and commercial Varnish-Plus. Collection is varnishstat-based at a 10-second default interval, including per-backend and per-storage metrics. The one gap to know about: no alerts ship configured for the Varnish integration, so you define your own rules.

What is varnishstat and how do I use it?

How much does Varnish monitoring cost?

Varnish monitoring can be free with open-source self-hosted tools or carry per-host, per-GB, or per-node charges on commercial platforms, depending on the tool. We do not quote list prices because they change and negotiated rates vary, but the pricing shapes differ significantly: Netdata charges per node with unlimited metrics. Datadog meters per host plus per-GB plus per-metric. New Relic charges per-GB ingest plus per user. Grafana Cloud meters per active series plus per-GB logs. Checkmk prices per monitored service; LogicMonitor and Site24x7 per device or monitor. Open-source self-hosted options like Zabbix, Icinga, and Checkmk Community carry no license fee but real operational cost. Check each vendor’s pricing page linked above for current figures.

Can I monitor multiple Varnish instances with one tool?

Yes, though the mechanics differ. Netdata supports multiple instances and remote instances within a single collector configuration. The Prometheus varnish exporter needs one exporter instance per Varnish instance, selected with the -n flag. Zabbix community templates handle multiple instances through agent user parameters. Icinga apply rules can monitor multiple Varnish instances with check_varnish.

The 10 best Varnish Cache monitoring tools, ranked

How we evaluated Varnish monitoring tools

Netdata

Grafana Cloud + Prometheus

Datadog

New Relic

Checkmk

Zabbix

Dynatrace

Site24x7

LogicMonitor

Icinga

Frequently asked questions

Book Your Free Demo