What if the most expensive container security tool in your stack is the one you got for free? That sounds like a riddle, but it plays out in real engineering teams every quarter. A free, open-source scanner racks up hours of CI compute and triage labour. A quote-based platform subscription swallows a budget line nobody can explain to finance. And the “cheap” cloud-native option quietly grows every time someone scales a node pool. Pricing in container security is genuinely confusing, because vendors are not selling the same thing.
This comparison breaks down how the main pricing models work, what each category of tool actually charges for, and where the hidden costs hide. No single “cheapest vendor” answer exists, but a clear map of the terrain does.
Why Container Security Pricing Is So Hard to Compare
Container security is not one product. It splits into at least four layers, and each layer has its own billing logic:
- Image scanning, which checks container images for known vulnerabilities before deployment.
- Runtime protection, which watches what containers do while they run and flags suspicious behaviour.
- Kubernetes posture management (CSPM/KSPM), which checks cluster and cloud configuration against hardening standards.
- Supply-chain and SBOM tooling, which tracks what went into a build and produces software bills of materials.
Vendors price across these layers in wildly different units: per node, per workload, per image scanned, per cluster, per developer seat, or per vCPU-hour. A “price per container” comparison is therefore rarely apples to apples. Two vendors quoting similar annual totals can be metering completely different things, which means your bill will diverge sharply as your environment changes shape.
The practical lesson: before comparing numbers, decide which layers you actually need. Most teams need scanning plus some runtime visibility. Far fewer need full posture management on day one. Buying the platform because it demos well, then discovering you are paying for three layers you do not use, is the single most common budgeting mistake in this market.
The Four Main Pricing Models, Side by Side

Here is how the models differ in practice, including how each one behaves as your environment grows.
| Pricing model | Typical unit | Scales with | Predictability | Common gotcha |
|---|---|---|---|---|
| Open source | No licence fee | Your team’s time and CI compute | High in cash terms, low in effort terms | Triage labour and database upkeep are real costs |
| Per node or per cluster | Each node or cluster | Infrastructure size | Medium | Node churn in autoscaling clusters can inflate counts |
| Per workload or per container | Each protected workload | Deployment count | Low to medium | Bursty deployments and short-lived jobs skew bills |
| Consumption-based (cloud-native) | Per resource or per vCPU-hour | Compute hours and resources | Low | Cost grows silently with scale and idle capacity |
| Platform subscription | Quote-based bundle | Negotiated scope | High once signed | Bundled features you never switch on |
Notice that only two rows are truly predictable: open source and a negotiated platform contract. Everything metered by workload or compute hour moves with your architecture, which is fine if your platform team watches spend and painful if nobody does.
Open-Source Scanners: Free Licence, Real Costs
Tools like Trivy, Grype, and Clair carry no licence fee, and all three are released under permissive open-source licences. That makes them the default starting point for most teams, and rightly so. A well-run open-source scanner pipeline catches the majority of known vulnerabilities before an image ever reaches a registry.
But “free” deserves scrutiny. The costs that do not appear on an invoice include:
- CI compute. Every scan run consumes build minutes. Multiply by every commit on every branch and the number becomes material.
- Vulnerability database maintenance. Scanners depend on continuously updated vulnerability data. Someone has to keep that pipeline healthy and current.
- Triage labour. This is the big one. A scanner that reports 400 findings per image is technically working and practically useless. Engineers must tune policies, suppress noise, and decide what actually blocks a release.
- Integration and glue code. Wiring scanners into admission control, ticketing, and reporting is engineering work that never appears in a licence comparison.
A useful rule of thumb: budget open-source scanning as a fraction of an engineer’s time plus CI spend, not as zero. Teams that do this often find the honest total is closer to a modest commercial tier than they expected, while teams that ignore it are surprised when triage becomes a full-time job.
Commercial Platforms: Why List Prices Barely Exist
Aqua Security, Sysdig, Prisma Cloud, Wiz, and Snyk all sell platform subscriptions rather than standalone scanners. Their public pages rarely show a number, because pricing depends on node counts, cloud accounts, modules selected, and contract length. This is not evasion for its own sake. Bundling scanning with runtime protection and compliance reporting creates genuine value, and the vendor needs a conversation to scope it.
For buyers, the practical consequences are:
- Quotes are the only real prices. Treat any figure you see outside a quote as indicative at best.
- Module selection drives cost more than volume. Adding a compliance module can matter more than doubling node count.
- Discounts reward commitment. Multi-year terms and larger scopes move the number substantially, which means the first quote is rarely the last.
- Bundling hides the per-layer price. You may never learn what scanning alone costs, which makes it hard to judge whether you over-bought.
A sensible tactic is to ask every platform vendor for a phased quote: scanning only in year one, runtime added in year two. The delta tells you what the extra layer really costs, and it gives you a fallback position if budgets tighten.
Cloud-Native Options: Convenient but Consumption-Shaped
Cloud providers sell container security as part of their own platforms. AWS Inspector and Microsoft Defender for Containers are the best-known examples, and both bill on consumption, per resource or per vCPU-hour, rather than per seat. The appeal is obvious: no new vendor, no procurement cycle, native integration with the cloud account you already run.
The trade-off is that cost scales with your infrastructure, not your team size. That cuts both ways. A small, efficient cluster can be remarkably cheap to protect. A sprawling environment with idle capacity, over-provisioned node pools, and short-lived jobs can generate a bill that grows without anyone approving a change. Consumption pricing rewards disciplined platform hygiene and punishes drift.
Two practical checks before committing: model the bill at your current scale and at double your current scale, and confirm whether the meter counts provisioned capacity or actual usage. The difference between those two can be large in autoscaling clusters.
Hidden Costs Nobody Puts in the Comparison Table
The line items that decide whether a tool is worth its price rarely appear in a pricing table. Watch for these:
Triage and alert fatigue. A tool that generates noise costs more than it saves. The best-value purchase is often the one with the tightest default policies, not the lowest quote.
Integration effort. Every platform needs wiring into CI, registries, admission control, and ticketing. Ask vendors to show reference integrations rather than promising them.
Compliance reporting depth. Paid tiers increasingly gate benchmark reporting. Configuration standards such as the CIS Benchmarks define what “compliant” means, and if your auditors expect benchmark evidence, the tier that produces it is not optional. The same applies to framework reporting for regulated industries.
SBOM and supply-chain features. Demand for software bills of materials rose sharply after US Executive Order 14028 in 2021 pushed supply-chain transparency into procurement conversations. Many vendors responded by moving SBOM generation into higher tiers, so check which tier actually includes it.
Training and onboarding. A platform your team cannot operate confidently is a sunk cost. Factor in ramp-up time.
A Practical Framework for Choosing
Rather than chasing the lowest number, run a structured comparison. This sequence works well in practice:
- Inventory what you already have. Native Kubernetes controls such as RBAC, Pod Security Standards, and admission policies cover more than many teams realise. Do not pay to duplicate them.
- Map your layers. Decide which of the four layers you need now and which can wait a year.
- Fix your growth assumption. Get quotes modelled at today’s scale and at a realistic twelve-month scale. Ask which meter moves.
- Price the free option honestly. Estimate triage hours and CI minutes for an open-source stack and put a number on them.
- Run a two-week trial with real images. Measure findings volume and false-positive rate, not feature lists.
- Negotiate the phase-in. Ask for year-one scope that matches year-one needs.
The teams that get this right tend to start open source, prove the workflow, and then buy a platform when triage volume or compliance requirements justify it. The teams that struggle usually buy the biggest bundle first and spend the next year explaining the renewal.
Conclusion
Container security pricing resists simple comparison because vendors are selling different things in different units. Open-source scanners are free to licence and expensive in labour. Commercial platforms are quote-based and reward careful scoping. Cloud-native tools are convenient and scale with your infrastructure whether you plan for it or not. The winning approach is not finding the cheapest line item. It is matching the pricing model to how your environment actually grows, pricing the hidden costs honestly, and buying layers in the order you need them. Do that, and the free tool stops being the most expensive one in the stack.
Frequently Asked Questions
Are open-source container scanners really free?
The licence is free, and tools like Trivy, Grype, and Clair carry no fees. The real cost is operational: CI compute for every scan, maintenance of the vulnerability database pipeline, and engineering time to triage findings and tune policies. Teams that price those hours honestly often find open source costs less than a commercial tier but is far from zero.
Why don’t commercial vendors publish their prices?
Most platform vendors bundle scanning, runtime protection, and compliance reporting into one subscription, so the price depends on node counts, cloud accounts, modules selected, and contract length. A public list price would mislead more than it would help. The practical answer is to request phased quotes so you can see what each layer costs separately.
Is per-node or per-workload pricing better?
It depends on your architecture. Per-node pricing suits stable clusters with predictable node counts. Per-workload pricing suits teams that want to protect specific applications regardless of infrastructure. Autoscaling clusters can inflate node counts through churn, while bursty deployment patterns can inflate workload counts. Model both against your real usage before deciding.
How does cloud-native container security pricing work?
Cloud provider tools such as AWS Inspector and Microsoft Defender for Containers typically bill on consumption, per resource or per vCPU-hour, rather than per seat. That makes them cheap at small scale and potentially expensive as infrastructure grows. Always confirm whether the meter counts provisioned capacity or actual usage, and model the bill at double your current scale.
Do I need a commercial platform if I already scan images?
Not necessarily. Native Kubernetes controls cover a lot of ground, and a well-tuned open-source scanner handles most pre-deployment checks. A platform starts to justify its price when you need runtime detection, compliance evidence against recognised benchmarks, or supply-chain reporting at a depth your in-house pipeline cannot sustain.
What should I check before signing a multi-year contract?
Confirm exactly which modules are included at each tier, especially compliance reporting and SBOM features, which are often gated. Ask for pricing at your projected twelve-month scale, clarify which meter drives cost increases, and request a phased scope so year one matches year one needs. The first quote is rarely the final one.