This guide is for a CTO, head of engineering or founder about to hand a live platform to an outside team. It sits in Netbase's cloud platform engineering cluster; the delivery side of the same work is described on the DevOps and platform engineering service page.
In this guide
- Why is a tool list not proof of operating production?
- Which questions should you ask a DevOps partner, and what should each answer show?
- What are the red flags, and what should you do about each?
- Which kind of DevOps partner fits the job?
- When should you not hire a DevOps partner?
- How do you run the selection?
- Where AI fits in a DevOps partner's work
- What delivery record exists, and what does not?
- Alternatives to a full partner selection
- Limits of this guide
- Frequently asked questions
- Next step
Why is a tool list not proof of operating production?
Most DevOps proposals list the same tools: a CI/CD service, infrastructure as code, containers, a monitoring stack. Installing them is a setup skill. Operating production is a different skill: noticing a problem before customers do, deciding at night whether to roll back, restoring data under pressure and writing down what changed afterwards.
A partner that has done this keeps records as a side effect. Pagers create rotas and alert histories, incidents create reviews, releases create pipeline logs and restores create dated test notes. The questions below ask for those records rather than for opinions, so a weak partner cannot answer them with a well-written slide.
Which questions should you ask a DevOps partner, and what should each answer show?
Ask every shortlisted partner the same twelve questions, grouped by production duty. A credible answer points to an artefact, with client details removed if needed; a weak answer describes intentions.
-
On-call
- Question
- Who is paged when a system you run fails at night, and how is the rota kept?
- A credible answer shows
- A rota, an escalation path and an alert history with acknowledgement times
-
Incidents
- Question
- Show one written incident review from the last year.
- A credible answer shows
- A timeline, impact, root causes and follow-up actions with owners, written without blame
-
Incidents
- Question
- What changed in your own process after that incident?
- A credible answer shows
- A named change to an alert, runbook, test or pipeline gate
-
Releases
- Question
- How does a change reach production, and who can approve it?
- A credible answer shows
- A pipeline run log with tests, approvals and the deploy step
-
Rollback
- Question
- Tell us about the last rollback you performed.
- A credible answer shows
- The release that failed, how it was detected and how long the rollback took
-
Restores
- Question
- When did you last restore a backup and check the data?
- A credible answer shows
- A dated restore-test note with what was checked and the time it took
-
Observability
- Question
- Which signals tell you a service is unhealthy before users do?
- A credible answer shows
- Dashboards and alert rules tied to a service objective, not only CPU graphs
-
Delivery health
- Question
- Which delivery numbers do you track for a client?
- A credible answer shows
- Trends such as deployment frequency, lead time, change failure rate and recovery time
-
Access
- Question
- How will your engineers access our accounts, and how is access removed?
- A credible answer shows
- Named identities, roles you create, MFA, short-lived credentials and an offboarding list
-
Ownership
- Question
- Who owns the cloud accounts, the code repositories and the infrastructure code?
- A credible answer shows
- Your company, stated in the contract
-
Handover
- Question
- What would we receive if we ended the engagement next month?
- A credible answer shows
- A handover checklist: runbooks, diagrams, credentials rotation, open-issue list
-
Scope
- Question
- What would you advise us not to do?
- A credible answer shows
- A specific item they would cut or defer, with a reason
The incident-review question matters most. Google's SRE book describes a postmortem as a written record of an incident, its impact, the actions taken, the root causes and the follow-up actions, and it argues for a blameless form; see postmortem culture. A partner that cannot show one, even redacted, has either not run production or has not learned from it.
For delivery health, the DORA metrics guide defines the measures most teams use. Ask for trends on a client system, not for a benchmark claim.
What are the red flags, and what should you do about each?
-
No written incident review can be shown, even redacted
- Severity
- Stop
- What to do
- Ask for a paid trial on a non-critical system first
-
"Our systems never go down"
- Severity
- Stop
- What to do
- Treat it as no operating record
-
The partner wants its own company to own the cloud accounts
- Severity
- Stop
- What to do
- Keep the root account, billing and code in your name
-
Shared logins or long-lived keys for engineers
- Severity
- Fix in contract
- What to do
- Require named identities, MFA and short-lived credentials
-
Write access to production requested in week one
- Severity
- Fix in contract
- What to do
- Start read-only; grant write access per agreed change
-
Backups described, but no restore test ever run
- Severity
- Fix in contract
- What to do
- Make a restore test an acceptance item
-
Only tool names when asked about results
- Severity
- Clarify
- What to do
- Ask which delivery numbers improved for a past client
-
The people on the call are not the people doing the work
- Severity
- Clarify
- What to do
- Meet the engineers who will hold the pager
The access rules in the table follow common cloud guidance. AWS's IAM security best practices recommend temporary credentials, MFA, least-privilege permissions and regular removal of unused users and credentials. The same principles apply on any provider.
Restore tests deserve their own line in the contract. AWS's Well-Architected guidance on backups lists "assuming that a backup exists" and restoring without checking the data among the common anti-patterns, and it recommends periodic recovery tests against the recovery time and recovery point objectives.
Which kind of DevOps partner fits the job?
| Option | Choose it when | Do not choose it when |
|---|---|---|
| Managed cloud provider | The platform is stable and you want routine operations run day to day | The platform still needs design or the product changes every week |
| DevOps boutique | One gap: a pipeline, infrastructure as code or a Kubernetes setup | Nobody inside your company will own the result afterwards |
| Product-and-platform partner | The product and the platform it runs on must change together | Only the infrastructure needs work and your product team is strong |
| In-house platform hire | Platform work is continuous and you can keep two people on call | You need the work done within weeks or cannot cover a rota alone |
Selection criteria across all four: who holds the pager after go-live, who owns the accounts and code, and whether the people interviewed are the people doing the work. A ranking of named companies on similar criteria is in the comparison of cloud consulting and DevOps companies.
When should you not hire a DevOps partner?
Do not hire one when the product runs well on a simple managed host and nobody complains about releases; the cost of another party in the loop will outweigh the gain. Do not hire one when the real problem is the application code, because a pipeline cannot fix slow queries or a tangled data model. And do not hire one when nobody in your company can own the cloud accounts and the bill after handover: the partner then becomes an owner by default.
How do you run the selection?
-
Name the job
Build a platform, migrate, automate releases or run day-to-day operations.
-
Score your own practices first
The DevOps maturity assessment scorecard shows which gaps you are hiring for.
-
Send the twelve questions in writing
Ask for artefacts, redacted where needed, before the first call.
-
Interview the engineers
Meet the people who will deploy and hold the pager, not only the sales lead.
-
Run a small paid piece of work
A pipeline or a restore test on a non-critical system shows how the team works.
-
Write operations into the contract
Response targets, access rules, a restore test and a handover checklist.
Where AI fits in a DevOps partner's work
AI assistants are now part of most delivery teams, so ask where they act and where an engineer decides. Netbase works with the major commercial and open-source models and chooses them per project. On cloud and DevOps work, AI can draft infrastructure code, pipeline steps, alert rules and a first incident timeline from logs. Engineers still review every change before release and own access design, secrets, the production deploy and the rollback decision.
Netbase has also delivered AI projects for clients that are not named, such as retrieval-augmented knowledge assistants, document AI and MLOps pipelines.
- In delivered work
Product recommendation engine
Built for 4over4's online printing store from browsing and purchase history; a delivered AI feature named with a client, on a store rather than a DevOps engagement.
- Available capability
AI agents in delivery
AI-assisted infrastructure code and pipeline review is a capability Netbase offers, not yet tied to a published DevOps case.
What delivery record exists, and what does not?
Netbase works as offshore development partner on a multi-tenant cloud ERP delivered as SaaS on AWS for a US client, not named, and builds and runs its own SaaS products, such as Cloodo Workspace. Netbase names AWS, Google Cloud, DigitalOcean and Cloudflare as the providers it uses and claims no cloud partner tier.
Its listed security practices include secure code review, role-based access control, MFA for admin dashboards, vulnerability scanning and disaster recovery. Netbase holds ISO 27001 certification and a SOC 2 Type II attestation for its own operations. It runs, maintains and upgrades what it builds under an agreed support scope, with response targets set per contract, and delivery runs in a six-step lifecycle with weekly reviews and KPI dashboards.
Most Netbase projects are delivered on fixed-price contracts; milestone-based and monthly team models are also offered. No uptime figure, recovery time, incident count or cost result is claimed on this page.
Alternatives to a full partner selection
A buyer with a strong in-house team can hire one senior platform engineer and use the twelve questions as an interview guide instead. A buyer with a narrow security worry can start from the security questions to ask a software development partner and add only the access and restore rows above. Choose the full selection when a partner will hold the pager.
Limits of this guide
- The questions fit web, SaaS and commerce platforms on a public cloud; regulated or embedded systems need more rows.
- A redacted artefact proves a practice existed, not that it will be repeated on your system.
- Severity labels are a starting point; your own risk tolerance decides the final call.
Plan the next step with a Netbase consultant
Frequently asked questions
Ask to see one written incident review from the last year, redacted if needed. It shows whether the partner has run production, how it explains failure and whether it changed anything afterwards.
No. A tier shows standing in a cloud vendor's programme. It does not show who answers an alert at night, whether restores are tested or who owns your accounts.
Not by default. Give named engineers roles you create, with MFA and short-lived credentials, start read-only and widen access per agreed change, and keep a list for removing it.
By showing redacted artefacts: a rota, an incident review, a pipeline log and a restore-test note with client names removed. Confidentiality hides the client, not the practice.
Next step
If you want a partner to show this evidence for a platform you are about to hand over, book a solution review with Netbase JSC's CEO, or read how Netbase runs cloud migration and modernization.
Related services and solutions
AI-ready cloud platform engineering for SaaS and commerce workloads Netbase provides cloud platform engineering for SaaS and commerce teams that need a secure, repeatable home for their software and its AI features, on AWS, Google Cloud, DigitalOcean or Cloudflare. We design the architecture, write the environments as code, connect managed and AI services and hand over a platform your team can run, built the way we run ours.
Learn More
AI-assisted cloud migration with a cut-over plan and a way back Netbase provides cloud migration services for companies running business applications on ageing servers or hosting to move them to AWS, Google Cloud, DigitalOcean or Cloudflare without losing data or control. AI speeds up the code and dependency assessment; every migration still has an inventory, a rehearsed cut-over, cost controls from day one and a tested rollback point.
Learn More
AI-assisted DevOps and platform engineering for routine releases Netbase provides DevOps and platform engineering for SaaS and commerce teams to make releases routine instead of risky. We set up CI/CD pipelines, repeatable environments and observability on AWS, Google Cloud, DigitalOcean or Cloudflare, with AI assistance for tests, reviews and incident triage, so every change ships the same way and problems surface before customers notice.
Learn More
SaaS product accelerator: launch an AI-ready SaaS on proven Netbase modules A SaaS product accelerator is a set of reusable Netbase modules for accounts, billing, roles and integrations that helps founders and product teams launch subscription software faster, with room for in-product AI from the first release. Reusing these modules can shorten development time, and the approach is proven on Printcart, Netbase's own web-to-print SaaS.
Learn More
Work with Netbase remotely
Every Netbase engagement is delivered remotely, so clients in any country and any industry can work with us; Netbase has worked with clients in 80 countries. Contracts and invoices are issued by Netbase JSC, and payment is made by bank transfer to Netbase JSC's company account or PayPal Business. Talk to the CEO, Huy Nguyen (David), on WhatsApp or by email.
Discuss a project
Netbase JSC helps organizations design, build, modernize, and operate digital products and AI-enabled business systems.project@netbasejsc.com
+84 937 869 689
91 Nguyen Chi Thanh, Dong Da, Hanoi, Vietnam
Get in touch
Tell us what you want to build, modernize, or operate.