FAQ
OpsCanvas is a platform which correlates cloud resources and related data from across your disconnected systems including your cloud account, CI/CD pipeline, Kubernetes clusters, and any other systems of record into a centralized, accessible intelligence layer. The intelligence layer powers a live feed of changes in your cloud, maps your resources by application, environment, and team, identifies ways to reduce your spend, and can give you answers to anything you need to know about your cloud in seconds.
OpsCanvas aims to solve tool fatigue which is the result of silos built up over time in the cloud, where DevOps, FinOps, SecOps, and DevSecOps often have their own teams, tools and areas of focus which has caused this tool proliferation. OpsCanvas aggregates and correlates the data from these systems centrally, producing a source of truth and the context required for faster decision-making and huge improvements in operational efficiency.
Most companies running software in the cloud have a limited view because there are multiple, if not many, systems of record which store different information related to your cloud. Cloud costs are stored with the cloud provider or hyperscaler, deployment histories and resource relationships live in the C/CD deployment pipelines, container usage and allocation is buried in Kubernetes logs and helm charts, and incidents are in yet another system. Since this information is constantly changing, companies burden engineers (DevOps, SRE, and software) with a steady stream of questions, report requests, and tasks which are time-consuming and a distraction. OpsCanvas eliminates the extra busywork to research, report, and stitch together data so teams can be empowered by having that information already correlated and at their fingertips.
There are two primary ways a company can generate a strong return on investment (ROI) using OpsCanvas: time and money. The clearest ROI companies produce is when they use OpsCanvas to remove unused or “no longer needed” cloud resources which we call ‘zombie waste’. This is estimated to be 30% of a company’s cloud bill on average, but it can be lower or higher depending on several factors including how long the company has been running their software in the cloud, how mature their processes are internally, and whether they are actively managing deployment pipelines or not.
The second potential ROI contributor is time saved, which can be as much or more in monetary value as the waste removed when accounting for the value of staff time. The time savings could include everything from time spent correlating data in Excel, deployment automations, time spent researching spikes in spend or other changes in the cloud, or producing chargeback spending reports and fulfilling other executive requests.
While the amount of cloud bill reduction varies for every company, the likelihood there is waste to remove and reduce your cloud bill is very high. OpsCanvas provides a company the best chance of not just identifying cost reductions, but also removing them because the context the platform provides makes decision-making easier and faster – often waste goes untouched because teams can’t come to an agreement, or the risk of ‘breaking something’ is too high due to the lack of complete information, or context, to make a confident decision.
We typically see a range of 10% – 30% of monthly cloud spend which can be removed, and the average waste for a company using OpsCanvas is 22%. There are a lot of factors which contribute to the end result of how much waste your company has accumulated over time.
No. While OpsCanvas has the automations to remove resources, we maintain a core tenant to keep the “human in the loop” until they ask to be removed and hand off to an agent. In addition, the platform has mechanisms in place to limit not only agents, but also certain user roles or interfaces from doing more than they should. The delete resource function is only available in the CLI, and only if a user has the right permissions. Agents are their own role, and would only be given authority for functions the customer assigns.
Yes, OpsCanvas is easy to install across all of your cloud accounts and CI/CD pipelines and supports having more than one provider in either case.
We have seen an average of 22% zombie waste across customers looking to remove waste.
No. A core design principle for OpsCanvas was to ensure companies would not have to spin up an engineering project to set up, and that the existing workflows, tooling, and governance in place would not need to change in order to use OpsCanvas.
All of the above. OpsCanvas recognizes that individual contributors, managers, and executives all have different needs and preferences for how they interact with the information, recommendations, and reports which are produced. The web app is ideal for seeing summaries of activity and reports, while the CLI is a natural fit for initiating actions, doing research, and generally staying in the flow with their normal day-to-day activities.
OpsCanvas is designed with a security-first architecture. We only ingest the metadata and signals required to understand your infrastructure and operations, not application payloads or customer data. All data is encrypted in transit and at rest, scoped by strict access controls, and logically isolated per customer. Our AI operates within your context and is anonymous, meaning it reasons over your infrastructure graph and decisions without training on or leaking data across customers or to an LLM.
Vibe-coding works great for simple IaC until the system gets complex, shared, or long-lived, not to mention the risks a lack of guardrails introduces. OpsCanvas provides guardrails, not to mention pulls it all together with context so your deployed resources have meaning and can be more easily managed without driving up costs or producing security weaknesses. For example:
- Why something was deployed
- What depends on it
- Whether it’s still needed
- What happens if you remove it
OpsCanvas adds organizational memory and intelligence to your infrastructure, so decisions made today don’t become mysteries six months from now.
The intelligence layer is OpsCanvas’ context and reasoning engine powered by the Cloud Intelligence Graph (CIG). . It correlates your infrastructure data, cost data, configurations, and decisions, turning raw data into signals and contextual understanding.
Instead of just showing resources, the intelligence layer understands:
- Relationships
- Ownership
- Intent
- Risk
- Cost impact
This provides humans and agents the proper context to make confident decisions, and will power the transition to Agentic AI operations when a customer is ready.
Context is the collection of data which informs better, more confident decisions for humans and agents alike. Ecosystems like cloud operations can be limited by lack of context due to different pieces of information required to understand a situation or make decisions often reside in separate systems of record or communications between teams, therefore slowing down decisions or disabling teams from ever making one.
A context graph is a data store that allows operators (AI Agents & Humans) to query for decisions about a certain action, decision, or correlated entity in an organization’s workflow. The graph stores Decision Traces, the 5 Ws associated with a particular action. The context graph is a machine encoded representation of institutional knowledge.
The Cloud Intelligence Graph is a proprietary implementation of a context graph for the OpsCanvas platform.
Yes. You can query the CIG with Oscar, our CloudOps agent to answer questions like:
- “What resources are costing money but have no owner?”
- “What depends on this service?”
- “What can I safely shut down?”
While the CIG powers the live operational feed, cost reduction recommendations, and resource organization, its purpose is to be accessible by agents and humans to have a complete, contextual understanding of anything happening in your cloud.
OpsCanvas is installed using lightweight, read-only installation of ‘recorders’ in your cloud accounts, CI/CD pipelines, and anywhere metadata can be captured to enhance the understanding of your cloud.
There is a recorder which gets installed in your cloud account and is read-only, but the core OpsCanvas application, web interface, and CLI all run in OpsCanvas’ own cloud account.
OpsRunner is the name of the recorder installed in CI/CD pipelines to gather deployment details for the CIG. Once in the pipelines, OpsRunner can be setup to automate common deployment activities.
This is the name of OpsCanvas’ waste detection system, which is able to scan a cloud account to identify unused or no-longer-needed zombie waste within 24 hours. The scan also has the option of flagging zombies for delete, which can then be done from OpsCanvas’ CLI.ud account and is read-only, but the core OpsCanvas application, web interface, and CLI all run in OpsCanvas’ own cloud account.
Waste is an ongoing challenge for companies, even those with sound practices in place. It’s estimated an additional 1-3% of new waste is added to a cloud account each month, which can quickly go from 0% back to 20% in a year or less.
The cost management tools from cloud providers are limited in what insights they can provide and are mostly difficult to use. They’re limited by not having context which means most of their recommendations need a deeper level of analysis before a decision can be made.
Traditional FinOps tools typically rely on tags and are not able to take actions. Tags are used to try to re-create resource dependencies, but they are limited in what they can capture and tags typically only cover 70-80% of cloud resources which means they have incomplete information. When you don’t have any context or incomplete information, a significant amount of time ends up being wasted on reviewing recommendations for false positives to ensure vital resources aren’t being removed.
OpsCanvas has a security-first approach, which means there is no need to share credentials and any sensitive information is never stored.
OpsCanvas installs non-invasive, read-only recorders in the workstream of critical systems of record and shares the metadata about the resources and related activities with the CIG.
Our Q&A agent, Oscar, is available and capable of answering questions and taking limited actions.
Agents operate within a set of guardrails and can only perform actions which they’re able to do 100% accurately. The context graph provides the additional information necessary to make good decisions with the proper context.
Decisions are a critical component of having durable context for your team, as well as agents.
OpsCanvas is built for platform teams, DevOps, SREs, FinOps, and engineering leaders who manage complex, shared cloud environments.
Most teams see immediate visibility within hours and meaningful cost and risk insights within days.
Yes, by default. Automation is opt-in and fully controlled.
Only when explicitly enabled and approved. You control when and how OpsCanvas executes actions.
The context graph provides a clear, auditable view of resources, ownership, and decisions over time.
Absolutely. Role-based access controls ensure users only see and act on what they’re authorized to.
Ownership is a first-class concept in the graph, not a tag buried in a spreadsheet.
Yes. Understanding dependencies and historical decisions makes root cause analysis faster and more accurate.
Your infrastructure remains unchanged. OpsCanvas does not create lock-in or hidden dependencies.
No. OpsCanvas complements IaC by providing context, understanding, and operational intelligence beyond code.
The graph-based architecture is designed to scale with large, complex environments.
Yes. By capturing intent and continuously monitoring context, OpsCanvas helps stop zombies before they’re born.
To provide agentic cloud operations skills focused on delivering measurable, positive outcomes for users. By being grounded on a source of truth, OpsCanvas is positioned to provide the hightest level of trust which we believe is a requirement of any agentic systems, especially one focused on assisting in the management of cloud operations.