About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.
When a business has an IT partner looking after its staff and computers and a separate provider hosting its servers, faults land in the gap between them unless the split is written down. Write a three-column table of responsibilities, one column for each party including your business, covering who watches, who acts first, who decides and who is told. Give the IT partner only the access its role needs, under named accounts, and share the table with both providers so they agree to it.
Why it matters
Three-party arrangements fail quietly. The hosting provider sees the server is running and assumes the application problem belongs to the IT partner; the IT partner assumes the slow response is the provider's network. Meanwhile staff wait. A shared table turns "not ours" into "theirs, and here is how to reach them".
How it works
Australian guidance on outside IT providers
The Australian Signals Directorate (ASD) publishes guidance for organisations engaging managed service providers. Its key actions include: "Include and enforce clear security requirements in contracts, such as access controls, monitoring and incident response obligations." and "Limit privileged access and give providers only the access they need to perform their role."
On accounts, ASD advises: "Accounts should be attributable to the MSP to enable easy identification of MSP activity in privilege allocation and logs." On timing, it suggests to "consider blocking MSP access by default and allowing remote access at an agreed time" and to "Correlate this access with a specific job ticket." On sign-in, after advising that sensitive administration be limited to particular provider staff with restricted access to critical servers, it adds: "Combine this with multi-factor authentication to limit malicious actors' ability to compromise critical assets."
On oversight, ASD says: "Logs should be stored in a centralised location." It adds that "A security administrator or independent party without access rights/accounts should regularly review logs for suspicious activity in the reviewed systems."
In a three-party arrangement, both outside parties may hold privileged access, so this guidance applies to each of them. The page was last reviewed in April 2023; it is still the published guidance, but check it for updates. Our own observation, not ASD's: where each party keeps its own logs, each can usually provide evidence about activity at its layer, which helps when the cause of a fault is disputed.
Layers decide first responders
A practical split follows layers. The hosting provider looks after the platform: physical hardware, the virtualisation layer, storage, the provider's network and, where contracted, backups and monitoring of the servers. The IT partner looks after what runs inside the servers and on staff computers: Windows settings, accounts, the business application, printers and office networks. Some services, such as operating-system patching, could sit with either; the table must say which.
What to choose and why
A three-column table
| Area | Your business | IT partner | Hosting provider |
|---|---|---|---|
| Approving access and changes | Decides | Requests | Requests |
| Server platform, storage, provider network | Told of incidents | Reports suspected faults | Watches and acts first |
| Windows updates on servers | Approves schedule | Acts first, or Not involved: the contract names one | Acts first, or Not involved: the contract names one |
| Business application | Owns vendor relationship | Acts first | Provides platform evidence |
| Backups and restore tests | Approves restore | Requests restores | Acts first: runs backups and restores (if contracted) |
| Staff computers and office network | Told | Acts first | Not involved |
| Security incident | Decides, with advice, whether anyone must be notified | Acts first: coordinates the response and investigates its layer | Investigates its layer and reports to both |
The table is a template; your contracts decide each cell. When you fill it in, mark exactly one "acts first" for each row. Where the template offers two options, as for Windows updates, choose one and delete the other. The security incident row shows one possible choice; some businesses name the hosting provider instead, or an incident response firm.
Access for the IT partner
- Named accounts for each technician, recorded in your access register.
- Rights limited to the servers and functions the partner supports.
- Remote access at agreed times, linked to a ticket, as ASD suggests.
- Multi-factor authentication on every partner account that reaches your servers.
- Removal on the last day of the contract.
Escalation between providers
Agree that either outside party can contact the other directly for a fault, with your business copied, so a ticket does not wait for you to relay messages. Record each party's support contacts and hours in the table.
Illustrative example
A 25-person firm's application slows every afternoon. The IT partner checks the application and finds nothing; under the shared table it reports a suspected platform fault to the hosting provider, copying the firm. The provider's monitoring shows storage latency rising during a nightly job that had moved to the afternoon. The provider reschedules the job, and both parties note the cause in their tickets. The example is illustrative.
Read next
Related articles you may want next.
- HostingOutside IT partners and application vendors in your hosting portal: time-limited access and reviewControl outside IT partner and application vendor access to hosted systems. Use named accounts, limited permissions, agreed end dates and an access register.Read article
- HostingManaged, co-managed or self-managed hosting: choosing how much of the server you runHow a business decides which layers of a hosted server it hands to a provider and which it keeps, using staff time, the need for administrator access and the evidence it must produce.Read article
- HostingProvider administrator access in managed hosting: who can log in, how it is logged, and break-glass accountsWhat administrator access a managed hosting provider holds to a customer's servers, what Australian and Microsoft guidance says about controlling and logging privileged access, how emergency break-glass accounts should work, and the questions to ask before signing.Read article
Risks and limits
- Contracts override tables. If the table and a contract disagree, the contract applies; update one to match the other.
- Shared credentials break accountability. Avoid partner logins shared between technicians.
- Gaps hide in rare events. Test the table with one scenario a year, such as a full server restore. Our companion article on who can ask for a restore covers how to authorise those requests.
- Notification has legal tests. Whether a security incident must be reported to affected people or a regulator depends on the law that applies to your business; agree in advance who will give you that advice.
- This is not legal advice. Liability and insurance terms between three parties need professional review.
Sources
Opened 15 November 2026; reopened 16 November 2026.
- Australian Signals Directorate: How to manage your security when engaging a managed service provider, published 21 December 2018, last reviewed 5 April 2023.
Related service: Managed support.
Learning notes
- With an IT partner and a separate host, write one responsibility table that both outside parties accept.
- Mark exactly one party that acts first for each area.
- ASD: limit providers' privileged access to what their role needs.
- ASD: use attributable accounts, multi-factor authentication and time-limited remote access tied to a ticket.
- Let the two providers contact each other directly, with your business copied.
This article was originally published on .