6 ways to be smarter and more strategic about your CloudOps spend
Quick wins and strategic actions to optimise your cloud spend-to-outcome ratio
Managing cloud spend remains a top challenge for enterprises. According to Gartner, organisations with little or no cloud optimisation plans end up overspending on cloud services by up to 70%, without deriving the expected value from it.
Gartner also projects global spending on public cloud services to spike 20.4% this year, to a total of AUD$697.01 billion. It is more pertinent than ever to tightly optimise and get the most out of your CloudOps budget.
CloudFinOps is not just a cost management discipline
It spans cloud productivity and capability and provides a springboard for sustainable competitive advantage. Getting the most of your CloudOps budget (only paying for what you need and use), requires a whole-of-organisation remit and a strategic cloud maturity trajectory.
Know exactly where to curtail your CloudOps costs
In tandem with revenue, are profitability and cost control. Our Modern Cloud ‘Run’ team save our customers 20%, on average, on their Cloud usage costs. In this blog, I will outline six of our most effective (simple and easy) cost optimisation tactics to immediately reduce your cloud expenditure.
“The Cloud doesn’t guarantee cost savings. Despite reducing capital expenditure, service subscriptions may even increase operating expenditures. Executive leaders must plan and account for the differences to ensure savings and other cost benefits are realised and recognised wherever they occur in the organisation.”Gartner, ‘How to Realize Cost Savings After Migrating to the Cloud’
But first, understand where you are on the cloud maturity trajectory
Before pinpointing areas for cutting your run costs, it is crucial to know and have a salient understanding of how your company uses your cloud platform today, vis-à-vis your future needs and strategic technology ambitions. You need this understanding to accurately monitor and measure your progress as your business ramps up its cloud maturity, ideally to the stage where your environment is continually optimised, with services federated, interoperable and proactively maintained on an open platform.
So, prior to addressing some hygiene factors for cloud economics, know the following:
WHERE YOU ARE ON THE CLOUD MATURITY MODEL
Are you at Cloud Maturity Mode 0 (CMM0), bound by legacy applications on on-premises, dedicated infrastructure, or perhaps on a multi-cloud environment, already proficient with some tooling and integration for automated cloud usage?
WHAT YOUR PRIORITY TECH GOALS & STRATEGIES ARE
Common customer scenarios are:
- To be cloud-first (SaaS, PaaS, IaaS)
- Buy before build
- Replace on-premises, or complement it
- Datacentre exit
- Saving money (do you know if you are actually saving?)
The low hanging fruit
1. Step 0 – create observability of your cloud spend
You can’t save money if you don’t know, or can’t see, where it’s being spent. There are many tools for this: AWS Cost Explorer, and an array of third-party tooling, such as from Cloudcheckr, Cloudhealth, Flexera.
Strategically, your infrastructure and ops leads need to actively work with your Finance team. They need to jointly establish a cloud expenditure budget to remediate overruns, and determine longer term capacity constraints (looping back to budget setting). Regularly publish and disseminate reporting on your cloud spend. This will enable you to identify areas for improvement, but also to celebrate successes internally – really socialise and celebrate these. Cloud cost optimisation needs to be elevated to a whole-of-organisation priority, as it is the baseline for both profitability and performance.
2. Tag & track resources
All the above cloud monitoring tools rely on meaningful and consistent tagging of your resources to be truly effective. Why does having a tagging standard for cost-attribution save you money?
To control cloud costs, you need to know who within your teams own certain resources, and who has the authority to make changes. Tags are critical to understanding and pinpointing cloud costs, as they assign metadata to each cloud resource.
My quick tips for effective resource tagging:
- Define your MVP requirements up front (know how you want to allocate cost)
- Create a well-defined tagging policy for consistent usage throughout your organisation (e.g. identifiers for your specific business – project, application, cost centre, geography – whichever categorisation helps you get a clear picture of where your spend is coming from)
- Start simple and define a minimum workable set of tags to deliver on your requirements
- Incorporate tagging within provisioning tools and templates
- Use automation
- AWS organizations can enforce mandatory tagging via Service Control Policies (SCPs) – or use your infrastructure-as-code tooling
- Make your CloudOps team accountable for accurate tagging (incorporate this into their KPIs; create friendly competitions to motivate this best practice)
- Review regularly and iterate according to changing business needs
3. Monitor your costs so you can manage them
Do you have cost alerts set up, and are you actively monitoring and responding to them? Are you doing any anomaly detection (e.g. identifying cost aberrations), and responding adequately? Use the tooling listed in 1) to help with cost surveillance.
Tips for monitoring:
- Set up cost alerts in all accounts
- Monitor and action any overruns
- Alert on cost-spikes (AWS Cost Explorer can help with this)
- Do cost attribution/allocation optimisation
- Continuously review and update thresholds according to your usage – an alarm that’s set too low and will be routinely ignored and therefore isn’t serving as an alarm anymore
4. Know & delete obsolescence
Cloud resources are like running water. You wouldn’t leave the tap on while you’re brushing your teeth. You can easily cut costs by turning off the tap while you’re not using it.
Crucially – pay attention to your non-production footprint. How large is it compared to production (i.e. what percentage of your total spend is attributable to nonprod)? This would be the potential percentage savings you can garner immediately.
Tips for getting rid of obsolete costs:
- Clear out those sandpit accounts as often as you can (ideally run automated clean-up on schedule – this has the added benefit of prompting your teams to use Infrastructure as code, lest their work in progress be lost due to clean-up)
- Use AWS Instance Scheduler to shutdown environments when not required.
- Savings can be significant, so be creative with how this is presented to your leadership teams (e.g. visual representation of the savings – turning off 7pm – 6am and weekends = up to 66% saving!)
- Enforce sensible account limits and lock out expensive services/large instance types unless needed – this reduces the risk of dev teams making expensive mistakes
- Resources that are used regularly but intermittently (such as dev and test environments) could be automatically turned on and off at scheduled times to proactively reduce costs.For example, a virtual machine made available on a 10×5 basis and switched off at all other times (including vacation/sick leave) could attract as little as 27% of the cost of leaving the VM running 24/365 on PAYG pricing.
- Regular checks also help uncover cost saving opportunities from redundant or dormant cloud resources that are no longer needed and can be decommissioned.
- Diligence before acting on idle resources is essential, as they may be in an idle state for a reason (e.g. for high availability purposes).
- Build your own processes and automation to tackle this area, or leverage third-party cloud management tools to quickly gain sophisticated, automation and optimisation capabilities.
This is a logical, but often overlooked hygiene factor for cloud instances – only pay for what you need and use. Know what the benchmark (or best practice) running costs are for your organisation, based on your existing and extrapolated workloads.
Right-sizing – like cost optimisation in general – is an ongoing practice of continuous improvement. As your workloads and computing requirements evolve, ensure you periodically carry out rightsizing exercises, track the outcomes, and instil baseline Service Level Indicators (SLIs) and align these to your Service-Level Objectives (SLOs).
To strategically manage these initiatives, assign leadership and accountability within your team to proactively tap into opportunities to right-size.
Easy tactics for right-sizing:
- Prune storage costs by determining the appropriate storage class for your objects (so you don’t pay premium costs for infrequently accessed data)
- Understand the acceptable levels of performance for each application
- Use RDS for databases and consolidate instances where it makes sense to do so (I.e. don’t unintentionally compromise a solution’s availability by over-consolidating and creating shared dependencies that reduce overall availability – unless of course you understand this and are ok with it)’Ensure you are monitoring EC2 Memory in CloudWatch to extract better sizing-recommendations from the tooling
- Watch out for over-provisioned EBS volumes and be conscious in your choice to use Provisioned IOPS. Migration tools are common cause of overprovisioned EBS so be sure to check this.
6. Sign up for Savings Plans and get rewarded for your loyalty
Once you understand what needs to run when and how big it needs to be, look at AWS or Azure Savings Plans to significantly reduce your expenditure.
Savings Plans offer significant cost savings in exchange for a commitment to a consistent amount of usage for a 1- or 3-year team. For example, AWS Savings plans build on Reserved Instances, but give you greater flexibility to change instance types or even Regions in which they are applied.
Tips to leverage savings plans:
- AWS Cost Explorer will make suggestions based on your history of utilisation, e.g. the EC2 Instance Savings Plans offer up to 72% savings compared to on-demand pricing on your Amazon EC2 Instances usage.
- Get the right balance between RI and PAYG purchasing. RI can be more cost effective for workloads with predictable, long-term scaling requirements, whereas workloads that are unpredictable and more volatile may suit a PAYG usage model.
- Be proactive about managing your RI portfolio – extrapolate your ongoing requirements ahead of RIs expiring, so you can be on the front foot in strategically tailoring your Cloud Savings Plans.
You need dedicated resources to monitoring & controlling your cloud run costs.
Cloud FinOps is a discipline in its own right. Have someone specifically assigned with this remit, so you are constantly watching your costs and coming up with ways to optimise. This person, or team, will need to manage the following cycle:
- Measure and allocate costs
- Determine budgets and thresholds
- Optimise your utilisation
- Forecast and adjust strategy accordingly (taking advantage of savings)
- Rinse and repeat
How Versent Modern Run can help
We know cloud maturity models, best practices and public cloud stacks inside-out. Our team are specialists in architecting, integrating and running cloud solutions that are optimised and make the most of modern cloud capabilities.
We have facilitated: Mirvac in cutting the running costs of their AWS platform by 40%; Boral saving US$1 million on their annual CloudOps; and Land Services South Australia shrink their AWS bill by 72%.
To start saving and proving increased profitability to your business, reach out to us for:
- A Cloud Economics consultation with me, Technical Director of Versent Modern Run. In this one-hour consultation, we discuss your cloud adoption trajectory / maturity, technical objectives, and pinpoint areas of immediate cost savings for your organisation.
- A Well Architected Review focusing on the Cost Optimisation. This review session takes 2-3 hours, and assesses your architecture against the AWS best practice framework. The review and customised report are complimentary. You also have the opportunity to receive funding credits from AWS once the review is complete.
Start now with Step 0
My parting words are to firstly focus on knowing where your money is being siphoned. Observe.