This week on The Cloud Show #theCloudShow I interview a very good friend and Azure Architect extraordinaire Rik Hepworth @rikhepworth of Blackmarble. See him turn the table around on me to begin to ask ME questions! ;)

[Follow the show]

We spent the episode partially reminiscing about how the cloud started and grew, but more importantly we talk about how organizations should think when strategically deliberating how to approach the cloud. That and more on The Cloud Show!

Reach out to me if YOU want to be a guest on the Cloud Show, or if you know someone who would make a great guest star! Hope you enjoy the show!

By the way, I particularly like this frame from the show:


About the Show Star Rik Hepworth: Rik is a Chief Consulting Officer at Black Marble, He helps organizations large and small use Cloud better. Whether that is a new adoption where governance and management are key or an application modernization project where technology choice and team knowledge are important factors in success. Through his public speaking and community involvement, He is a recipient of the Microsoft MVP Programme award for his work with Azure. He is also proud to be a member of the Microsoft Regional Director Programme. As an event organiser, speaker, and lapsed blogger, he is keen to share his knowledge with the community and encourage & support others as they do the same. His grey hair is an indication of how long he has worked in IT, and in that time he's done most things you'd expect of an IT Pro: From crawling under floors pulling cables, to managing large-scale automated OS rollouts, to architecting solutions for on-premises and cloud. Now he spends more time working with his customers' management, sharing that breadth of experience and learning.

Comment Section

You can just “lay down your credit card” and start consuming Azure for your business. No problem, right? No, that’s not entirely right. It matters how you purchase Azure, and there are several options, but it matters more how you follow up on cost.

This question in the Cloud Clinic is one that will inevitably “hit” your company when the initial party mood can turn into a hangover. Hopefully, before it does, you have learned to reign in, control, and master cost management in the cloud. Who am I kidding, evidence, unfortunately, speaks to the contrary.

I have seen multiple companies take to the cloud in maybe ambitious strides, not really knowing exactly how to measure business value (yet), and then later realize that they have been haemorrhaging money too much for too long because of uncontrolled costs and a general free for all attitudes in the company.

Related to this question are the problems that arise from not having the right measurements and monitoring in place (Establishing and monitoring access to different environments (part 1)) and the conversation about how to approach a solution for this challenge (Establishing and monitoring access to different environments (part 2)).

Read on below and find out in this week’s episode!

The Cloud Clinic is a series on the #AzureEnablementShow where we focus on answering caller questions about using the cloud. It is difficult to start out right, and it is difficult to stay on an optimal path in the cloud journey. "I thought the cloud would be better than this, but I have some questions!" This is the show where you can have Your question answered! Please reach out to me on social channels, or comment here, or on YouTube, and we might be answering Your Cloud Clinic Question next!


Use tagging and cost management tools to keep your org accountable

Please enjoy this episode on YouTube:


Here is the episode on Microsoft Learn:

How can you tell if your cloud investment is a good one?

Oh that’s “simple”, you measure! If it truly was simple, then everyone would do it. The baseline is that you must have tags for cost management on your cost consuming Azure Resources. If not initially, then certainly later, you will want to be enabled to slice and dice various costs. How much are we spending on our production environments versus our test? The increase in cost over time, how much of it is accrued due to rising development costs? Could it be that we maybe are not cleaning up old development and test resources that we are no longer using? It that’s true you certainly would not be the first. It is important, over time, to have a check on which cost is what, and use those checks to work toward an automated and clean development and test landscape. It is also important to measure production cost versus revenue from those resources over time. Is there something that needs to be revisited in our architecture after we gained all these paying customers? Is the rising cost realistic compared to our growing revenue? What may be reviewed and optimized?

How can you associate the cost of a resource with a person or team?

Again, with the tags! It is a fundamental function of Azure to put metadata keys and values “tags” on all your resources. You can enforce this by policy and make sure it always happens. Azure is just now releasing the feature that allows you to easily and powerfully “Group and allocate costs using tag inheritance”. That is certainly worth looking at.

How does using Azure budgets add accountability?

When a system accountable person (or team) is assigned a budget for their Azure spend, their mindset shifts. I have seen so many cases where budgets are not used, and much later a very sour conversation about “how could it have cost this much” happens. In one case an Azure Enterprise Agreement of $1M was supposed to last for a year. After seven months, that money was all gone. The conversations that followed were not entirely friendly. That company did not use tagging the right way, and they did not use budgets. Internal post-mortem reviews revealed that technical employees felt outside of the business goals and that nobody listened to them. They in turn did not feel trusted and therefore not accountable. That cost hundreds of thousands of dollars in this case. They are most certainly not alone.

Remember, a budget does not have to be spot on correct to begin with. Just make sure to set one up, for example for a subscription, and then follow up on the cost. Is the cost reasonable? Does the budget need to be higher, or is there room to optimize and automize and reduce the budget?

Use tags and budgets and work to make technical ownership inclusive in the business, so that those that spend feel they are accountable for spend, and potentially rewarded for not overspending!

What are the first steps to get started?

Comment Section

DevSum conference 2023

Some conferences are special – the good Eurpean kind of special. DevSum falls in that category. I suppose a large part of the reason is that it is on home turf, in Stockholm, Sweden.

This year the conference content committee saw fit to select my talk on Four pre-flight checks for Azure Cloud. This is a particularly important topic these these days, and the audience frequently comment a lot of recognition. The fourth thing in my talk is “references”, but three large areas of items to cover _BEFORE_ you start are ensuring you have a good strategy, which leads to a good plan, and then to also ensure you empower and arrange your company organisation around cloud.


Technology may be advanced, but is frequently rather straigth forward. Humans, on the other hand, are _NEVER_ straight forward.

We tend to talk about and work on the tech, because that feels safe and usually not so contentious. As an industry we must talk a lot more about these other things, because that’s where the real cost of cloud hides, and this is where a quite signifficant portion of project time is spent!

In my talk I share a lot of real project stories (names have been changed to protect the victims), and a lot of personal experiences and tips earned in the trenches of cloud projects since the very dawn of the cloud!

Also at the conference, I met up with a ton of great friends and I saw multiple great talks delivered by some of those friends. Always phenominal to catch up both socially, and also professionally at conferences!


Thank you a lot DevSum for having me over – I appreciated this a lot! Hope to see you again next time! <3

Comment Section

Recently on the Azure Triumphs show, Johan Åhlen and I had a chance to share about how we approach setting up a productive, agile and effective Cloud Strategy for Region Skåne – one of the largest political geo-regions in Sweden, where we both live.

We talked about high and low, and mixed up a nice concoction of experience and advice for how to help a non-cloud public sector organisation embrace and adopt cloud. It is challenging, and quite rewarding, and it is a much more exiting journey than you might imagine. Johan is driven by improving healthcare for citizens of his region. I can only agree. It does feel warm and good to be able to help my region, where my family, relatives and a lot of people I care for work, live and raise their families. This is what it is all about, folks! Taking your experience and expertise and putting it to work to make a difference where you live!

This was part of the Azure Skåne #GlobalAzure 2023 event. Join the Azure community in Skåne and follow Global Azure for future #global #azure events! #mvpbuzz

Listen here:

Thank you very much Nikos Delis for having us on your show!

Comment Section

Last week on the Cloud Clinic we talked about what bad things can happen to your cloud when technical people have all the wrong access to your cloud assets. We’re talking costs running amok, instable work environment that are prone to environment drift and that cause your development to slow down to a crawl, and risk for mistakes that should not be possible to make. Those and more issues we list in that episode. Check it out: Establishing and monitoring access to different environments (part 1 of 2)

Now, let’s talk about the remedy! What do you do to avoid all the many problems that come from technical people having wrong access to the cloud?

Read on below and find out in this week’s episode!

The Cloud Clinic is a series on the #AzureEnablementShow where we focus on answering caller questions about using the cloud. It is difficult to start out right, and it is difficult to stay on an optimal path in the cloud journey. "I thought the cloud would be better than this, but I have some questions!" This is the show where you can have Your question answered! Please reach out to me on social channels, or comment here, or on YouTube, and we might be answering Your Cloud Clinic Question next!


Establishing and monitoring access to different environments (part 2)

Please enjoy this episode on YouTube:


Here is the episode on Microsoft Learn:

It is a remedy in two parts. And of course, you need to lay down a solid strategy and stay the course in practice in your company. This is comparable to finishing that course of penicillin once you feel the symptoms going away. If you don’t the illness will come back, but it is challenging to motivate continued cure when you don’t feel sick. Trust me says the doctor about penicillin. Trust me says I about the following advice! It is not all that easy to do, but the rewards are well worth it.

There are two remedies to consider: Role Based Access Control and Privileged Identity Management. The former is what the Japanese call “kata”. It’s just good form. The latter is a bit tricker to manage, but when well rolled out, it empowers your organisation so very much!

Role Based Access Control (RBAC)

The way you grant someone, or something (a system) access to resources in Azure is by means of Role Based Access Control (RBAC). The thing about this is that RBAC data is stored in Azure, while identities of the people accessing the cloud are stored in Azure Active Directory (AAD). This leads to inconsistent access that is difficult to manage in Azure.

What I propose you do instead is take advantage of the AAD for what it is good at – users and group membership. The AD has “security groups” that are used to group users together to grant them access to something. You put users as members in the group, and then you grant the group, which is just another AD object just like a user identity, access to the intended system!

This means, in Azure, you will see only groups having access. You can create and maintain these as standard groups and standard access using automation, and you can easily automatically audit these groups. This makes your cloud access management cleaner!

Back in your AAD, to grant a person access to a system in Azure is as easy as making them a member of the right security group. AD is GREAT with auditing and managing who is a member of what group. This is what the AD is built for!

The outcome is that you now play the strength of each system, rather than ending up in this very uncomfortable split that gives you an inevitable pain in the groin! Winking smile

But wait, theres’s more!

Use the AAD security group owner to delegate membership

To let group leaders in your projects, self-manage who are members in their teams, and consequently who has access to the systems in Azure they oversee, you can delegate ownership of AAD security groups to them!

Regular users in the AAD do not necessarily have the right to add or remove members from security groups. This means they must ask someone else to set up access, say when a new team member joins the team, or revoke access when one leaves. The access administrator role in AAD has this privilege, but that creates an inconvenient bottleneck and leads to an access request granting system that is not very quick and agile.

Instead, what you should do is consider granting security group owner privileges to your team leads. They will then be able to manage membership of the security group themselves! This affords them a limited right to decide who is in the team and so have access to the Azure resources. This little trick is so very valuable it should come in gold!

Privileged Identity Management (PIM) is the next level of access control

Privileged Identity Management is a premium Active Directory feature that allows you to preconfigure access that can be activated only when needed. The tough bits here are two. One is that you need to pay for the AAD P2 license. But again, trust the doctor, it is well worth the money if used properly. Two is that you must invest some effort into fully adopting the feature into your organisation.

It is a bit tricky to configure PIM and that’s why I see organisations struggling to adopt it. You should start with a POC for PIM, if you will, and enable it for one team all the way. Then you take those experiences to learn how to roll out in your company all the way!

PIM means, as alluded above, that you configure access as the potential to have access, but the user does not really have access right now. It is just-in-time activated by the user. The main power here is that, if a user does not have access to a system, they cannot inadvertently make mistakes and modify that system by accident.

Use multi-factor authentication to further strengthen security

Another great combination here is MFA, or Multi-Factor Authentication. This is a standard feature that means you must use your phone or similar in conjunction with your password to authenticate. It is a great security addon that almost entirely removes the risk of credentials being stolen.

That plus other AAD features such as geofencing which blocks sign-ins from users that “all of a sudden” are attempting to gain access from Russia och China when they were in the office five minutes ago.

Manager approval is another method for controlling access to production

Activation of your access using PIM can be set up in such a way that you have to give written justification for the elevation of privileges. You can have the process that a work item must be related to why the user needs access, and this may be audited after the fact.

For additional security, again a risk of bottlenecks, you can also activate manager approval for users’ PIM activations. Just make sure you don’t have that activated for the support team that would need to phone a manager in the middle of the night on a weekend to grant access during maybe a critical security incident. That’s not going to sit well with upper management. Use this power only where it makes sense, and make sure more than one manager can be on call to grant access!

Don’t forget to audit access and review if a person still needs that extra access

Having limited access also provides useful audit logs so that you can determine who had access to certain environments at specific times. Any time a person accesses a system in Azure it is audited in the Activity Log. Same goes for any time a user elevates access using PIM, and the AAD also audits when users are added to or removed from security groups. Feed all this data into your SEIM system and have it analysed there, for example using Azure sentinel, which is purposefully built for exactly this type of auditing and security management.


There are so very many things you can do to have a proper access control, and which brings you so many benefits in terms of saving time and money, and in mitigating risk of data loss and thwarting attacks. Use proper RBAC with AAD security groups and consider PIM! You will not regret that investment, because you end up saving a ton!


Comment Section