Skip to content
EncryptTechBD
Home Industries Our Work Blog Pricing About Contact Discuss Your Project
বাংলা (BN)
Cloud, DevOps & FinOps 6 min read

Container image security: what it costs in Bangladesh

A cost model for container image security in Bangladesh, covering scanning, triage, base image ownership, signing and the exceptions that never close.

E
EncryptTechBD · 0 comments
Server rack with networking equipment in a data room
Photo: Famartin / Wikimedia Commons

A container image security quote usually starts with a scanner and ends with a number. The scanner is the easy part. The expensive part is deciding which findings matter, rebuilding images without breaking the application, proving that a deployed image is the one that was tested, and keeping all of that true next month.

This guide is for a team in Bangladesh that already runs containers and needs to budget image security as an operating capability. It does not compare scanning products. It separates the cost components, shows how effort scales and provides a worksheet for planning.

What the spend covers

  • Image inventory. A record of every image, its source repository, owner, base image and deployment location.
  • Scanning. Vulnerability checks on the operating system packages and language dependencies inside each image, at build time and in the registry.
  • Triage. Deciding whether a finding is reachable, exploitable and relevant to the deployed configuration, then assigning a fix date.
  • Remediation. Updating base images, rebuilding, testing and redeploying affected applications.
  • Base image ownership. Maintaining a small set of approved bases, patching them on a schedule and retiring the ones nobody uses.
  • Supply-chain evidence. Software bills of materials, image signing and provenance records that show what was built and by whom.
  • Registry and admission control. Who can push, which tags are mutable, and which images are allowed to run.
  • Runtime detection and response. Monitoring for unexpected processes, file changes or network connections inside a running container.
  • Exception management. Documented acceptances with an owner, a reason and an expiry date.

Why the first quote is often too low

Scanning is usually priced per image, per repository or per compute unit. That is a predictable input. The unpredictable input is the number of findings the team must triage and the time needed to rebuild applications that have not been touched in months.

An application with an old base image and hundreds of transitive dependencies can generate a long list on the first scan. Many findings may be irrelevant to the deployed configuration, but someone still has to make and record that judgement. A proposal that excludes triage is selling the alarm, not the response.

What drives effort

  • Number of images and repositories. More images mean more scans, more owners and more rebuilds.
  • Base image diversity. Five approved bases are easier to patch than thirty variants inherited from different projects.
  • Rebuild frequency. Images rebuilt weekly absorb patches naturally. Images rebuilt twice a year turn every patch into a project.
  • Dependency age. Old frameworks and pinned libraries increase the remediation work and the chance that an upgrade breaks behaviour.
  • Registry count. Multiple registries and cloud accounts multiply access control and evidence work.
  • Gate strictness. A warning-only scan costs less to operate than a build-blocking policy, but it also changes behaviour less. Blocking requires an exception process that works.
  • Runtime scope. Detecting behaviour inside running containers requires an agent, a policy and someone to investigate alerts.
  • Evidence requirements. A customer or auditor may ask for SBOMs, signatures and proof that a deployed digest matches the approved build.

A formula and calculator logic

Score the environment before comparing proposals:

Effort score = images × base image churn × triage load × gate depth

  • Images. One for up to five, two for six to twenty, three for more than twenty or any estate with unknown images.
  • Base image churn. One for a single maintained base rebuilt monthly, two for a few bases on a quarterly cycle, three for many inherited bases with no clear owner.
  • Triage load. One when most findings are patched by a base update, two when application dependencies need upgrades, three when old frameworks require code changes.
  • Gate depth. One for reporting only, two for blocking on critical findings, three for blocking with signed provenance and an exception workflow.

The product is a relative size. It helps explain why two teams with the same number of images receive very different quotes.

Estimate the run cost with a second expression:

Run load = rebuilds per month × remediation time + open exceptions review

Rebuilds should be a scheduled habit, not a reaction to a scanner alert. A monthly rebuild cycle is easier to staff than an urgent rebuild during a campaign.

Budget worksheet

The shares below are planning proportions for an approved budget, not vendor rates.

Cost lineUnit to countPlanning shareOwner
Image inventory and ownershipImages, repositories, teams10 percentPlatform lead
Scanning and CI integrationPipelines, registries, images20 percentDevOps lead
Triage and remediationFindings, rebuilds, applications35 percentEngineering lead
Base image maintenanceApproved bases, patch cycles15 percentPlatform lead
Signing, SBOM and evidenceImages, attestations, audits10 percentSecurity lead
Contingency and trainingUnknown dependencies, team time10 percentProject sponsor

Questions to ask before approving a proposal

  • Does the scope include triage and remediation, or only scanning?
  • Which registries and pipelines are connected, and which are excluded?
  • Who owns the approved base images, and how often are they rebuilt?
  • What happens when a critical finding appears during a release freeze?
  • Are SBOMs generated and retained, and can they be exported?
  • Is signing enforced at admission, or only recorded?
  • How are exceptions approved, reviewed and closed?

A 90-day sequence

  1. Weeks 1–3: inventory images, owners and base images; connect the registry and run a baseline scan without blocking builds.
  2. Weeks 4–6: reduce to a small set of approved bases, fix the highest-risk findings and establish a rebuild schedule.
  3. Weeks 7–9: add a critical-severity gate and an exception workflow; generate SBOMs for every build.
  4. Weeks 10–12: add signing and admission checks, test a rollback, and hand over the monthly review.

Track scan coverage, the age of the oldest critical finding, rebuild lead time and the number of exceptions past expiry. A low finding count with no rebuild schedule is usually a sign that nobody is looking, not that the images are clean.

Our Managed Technology & Cybersecurity division builds container pipelines with scanning, signing and monitored runtimes. The pricing page explains how assessment, implementation and ongoing support are scoped. For the product comparison behind this decision, read container image security: five approaches compared.

Sources

Services that deliver this

Want this implemented properly?

Pricing →

Every engagement starts with a written scope. If a requirement is outside it, we quote it separately before starting.

Responses (0)

No comments yet. Add the first one below.

Leave a response

Your email address will not be published. Required fields are marked *.

Keep reading

More field notes