Skip to content

Independent marketplace for verified cloud accounts

The Post-Handover Security Checklist for a Cloud Account

A time-boxed checklist for the first hour, first day and first week after you take over a cloud account, covering access, billing, logging, network exposure and running resources.

The riskiest period in an account’s life is the few days after it changes hands. Someone else knows the history, the shortcuts and possibly the keys, and you know almost nothing. The fix is not cleverness, it is sequence. Do a small number of things in a fixed order and the unknowns shrink quickly.

This checklist is organised by time rather than by topic, because that is how handovers actually feel. Treat the first hour as containment, the first day as audit, and the first week as instrumentation. Nothing here requires tooling you do not already have in the provider’s console.

Before you change anything, take an inventory

Resist the urge to start rotating things immediately. Spend fifteen minutes recording what exists, because once you begin changing credentials you lose the ability to tell what was original. Capture the account identifier, the registered email, the list of projects or subscriptions, every region with resources in it, and the current month to date spend.

Pay particular attention to regions. Consoles default to one region and happily hide compute running elsewhere. Open the billing view that aggregates across regions, or the provider’s global resource search, and write down everything you find. An inherited instance in a region you never visit is the classic source of a bill nobody can explain.

Record the handover timestamp too, to the minute if you can. Every later question about whether an action was yours or someone else’s resolves against that line in the audit log. Without it you will find yourself squinting at events and guessing, which is a bad way to decide whether to panic.

The first hour: close the obvious doors

Containment comes first. Change the console password for the owner identity. Remove any multi-factor device you do not physically hold and enrol your own. Check for alternate contacts, secondary emails and recovery phone numbers, and strip out anything that is not yours.

Then look for additional human logins. Most consoles make it easy to list users but easy to forget federation. Check whether any external identity provider, SSO connection or linked account can mint a session. If one exists and you did not configure it, disable it until you understand it. The detailed treatment of root email, authentication factors and keys lives in the guide to securing an inherited account, which is worth reading in parallel.

The first day: audit identities and permissions

With the front door closed, work through who can do what. List every user, role, service principal and service account. For each one, ask a single question: if this identity were compromised right now, what could it reach? Anything that answers “everything” and is not operated by you should be removed or reduced the same day.

Look specifically for policies attached directly to identities rather than through groups, because those are usually ad hoc grants that outlived their purpose. Look for wildcard permissions. Look for roles that can assume other roles, since a narrow looking identity can be a path to a wide one. If least privilege is new to you, the fundamentals of least privilege covers the model properly.

The first day: cut off programmatic access you did not create

Human logins are the visible half. The other half is keys. Enumerate access keys, API tokens, personal access tokens, deploy keys and anything issued to an integration. Note the creation date and the last used timestamp for each. Keys that have never been used are easy decisions. Keys in active use need a replacement before removal, not after.

Check third party integrations too. Monitoring agents, backup tools, cost dashboards and CI systems all hold credentials into the account. If you do not recognise the integration, assume it belongs to the previous operator and remove it.

Secrets stored inside the account need their own pass. Look in the provider’s secret store, in configuration services, in container definitions, in startup scripts attached to instances, and in any build pipeline the account can reach. Credentials embedded in a machine image are the hardest to find and the easiest to forget, so note any custom images you inherited and plan to rebuild rather than reuse them.

The first week: billing alarms before anything else

Put a spending alarm in place early, even a crude one. A budget with a threshold and an email destination takes a few minutes and catches the failure mode that hurts most, which is silent spend. Set the threshold low enough that normal activity trips it in the first month, then raise it once you know the baseline.

Also check for commitments. Reserved instances, savings plans and committed use discounts are obligations that keep billing whether you use the capacity or not. They are not visible in the resource list, only in the billing console. If the account carries one you did not expect, factor it into what the account actually cost you. Why cloud bills surprise people covers the other common sources.

The first week: turn on logging and keep it

An account without an audit trail cannot be investigated. Enable the provider’s management event logging, send it somewhere outside the account if you can, and set a retention period you are comfortable defending. Do this before you need it, because logging is almost never retroactive.

Pair the trail with a small number of alerts that fire on things that should be rare: a new identity being created, a policy being attached, logging being disabled, a key being generated. Four alerts you read beat forty you ignore. Audit logging from day one has the longer version of this argument.

While you are in the logging settings, check whether logging was ever switched off. Most providers record the act of disabling a trail even when the trail itself goes quiet, so a gap in coverage is usually visible as a pair of events rather than as silence. A deliberate gap in an account’s history is a strong signal about how it was run, and it should change how much you trust the rest of your audit.

Network exposure you did not create

Inherited accounts collect open ports. Walk the firewall rules, security groups and network access lists in every region that has resources. Flag anything open to the whole internet on a management port. Flag anything with a rule described as temporary, because temporary rules outlive the people who wrote them with depressing reliability. Check for public IP addresses attached to machines that do not need them, and for storage buckets with public read enabled.

Also check load balancers and DNS records pointing at the account, since those can expose a service you have not noticed. If the account is on a provider you are still learning, product pages such as Buy Digital Ocean Account and Buy Vultr Account are a quick way to orient yourself on where each console puts its networking controls.

Running resources you did not ask for

Now deal with the resources themselves. Separate them into three buckets: things you want and will keep, things you do not want, and things you cannot identify. The third bucket is the interesting one. Do not delete it immediately. Stop it if stopping is reversible, tag it with the date, and give yourself a week to find out whether anything breaks.

Snapshots, images, idle volumes and orphaned load balancers are the usual residue. They bill quietly and nobody misses them. Clear them out once you are confident, and consider a tagging convention from the start so the next audit is faster. For a comparison of how different platforms present this, Hetzner against DigitalOcean shows two quite different console philosophies.

Write it down and set a review date

Finish by recording what you found and what you changed. A single page is enough: the inventory, the identities you removed, the keys you rotated, the alarms you created, and the open questions. This becomes the baseline you compare against next quarter, and it is also the document you hand over if the account ever moves on again.

Resist the temptation to make it elaborate. A checklist that fits on one screen gets used; a thirty page runbook gets skimmed once and then ignored. The value is in the repetition, not in the formatting. Keep it in the same place as your infrastructure code so it travels with the account rather than living in someone’s notes application.

Put a calendar reminder for thirty days out to repeat the identity and key audit. Most of what goes wrong in an inherited account goes wrong because the first pass was never followed by a second one. If the audit keeps producing surprises, the honest answer may be to migrate workloads into a fresh account instead, and the case for account isolation explains why that is often cheaper than it looks.

Related comparisons

AWS vs Vultr
vs

AWS vs Vultr

Some teams need many small locations and fast access to bare metal or GPU capacity more than they need a…

AWS vs Azure
vs

AWS vs Azure

Azure is usually chosen through identity and licensing rather than through infrastructure. This comparison covers enterprise agreements, hybrid estates, hierarchy…

AWS vs Hetzner
vs

AWS vs Hetzner

For steady always-on compute the price gap between Hetzner and AWS is wide. This comparison works out what that gap…

AWS vs Linode
vs

AWS vs Linode

Linode under Akamai is a different proposition than the independent host it was. This comparison looks at flat pricing, edge…

Accounts discussed in this guide

Buy AWS Account
Buy AWS Account
Amazon Web Services

Buy AWS Account

Amazon Web Services has become synonymous with cloud computing itself, and for millions of developers and businesses around the world, an AWS account is…

From$25 14 options

Keep reading

CI/CD Pipelines That Deploy to Cloud Infrastructure Safely
CI/CD Pipelines That Deploy to Cloud Infrastructure Safely
DevOps and Deployment 7 min read

CI/CD Pipelines That Deploy to Cloud Infrastructure Safely

A practical look at how to structure a deployment pipeline that touches real cloud infrastructure, with particular attention to how the pipeline…

IAM Fundamentals: Users, Roles and Least Privilege
IAM Fundamentals: Users, Roles and Least Privilege
Security and IAM 7 min read

IAM Fundamentals: Users, Roles and Least Privilege

Identity and access management decides who can do what in your cloud account. This guide covers users versus roles, how permission decisions…

Need the account, not just the theory?

Browse documented configurations across every provider in the catalog, or ask us which one fits your workload.

Telegram