This guide is for CTOs, heads of engineering and product owners who run a customer portal, an ERP, a commerce site or a SaaS product in production with a small team and no reliability function. It sits beside Netbase's cloud platform engineering service, where service level objectives are one of the operating rules Netbase writes down with the client.
In this guide
- What are SLOs and error budgets, in plain terms?
- Why do small teams avoid SLOs, and what does it cost them?
- Which SLOs should a business application start with?
- How do you use an error budget in release planning?
- Setting up SLOs with a small team
- When should you not set SLOs yet?
- Where AI fits in SLO work
- What delivery record exists, and what does not
- Alternatives: other ways to manage reliability
- What are the limits of this guide?
- Frequently asked questions
- Next step
What are SLOs and error budgets, in plain terms?
A service level indicator (SLI) is a measurement of what users experience, such as the share of requests that succeed or finish within a time limit. A service level objective (SLO) is the target for that indicator over a window, such as 99.5% of checkout requests succeeding over four weeks. A service level agreement (SLA) is the contract, with consequences, that sits on top.
The error budget is the gap between the SLO and perfection. Google's SRE Workbook gives the arithmetic: with a 99.9% success SLO, a service that receives three million requests over four weeks has a budget of three thousand failed requests in that window. An outage that fails half of them has spent half the budget.
The point of the budget is a decision rule. While budget remains, the team ships features at the normal pace. When it runs out, reliability work moves ahead of new features until the service is back within its objective.
Why do small teams avoid SLOs, and what does it cost them?
Small teams usually skip SLOs because the practice sounds like it needs an SRE department, a dedicated tool and a 24-hour rota. None of that is required. The core is a written target, a dashboard that shows it, and a short policy the product owner has signed.
Without that, reliability is decided by the loudest stakeholder. After an outage, everything stops for a hardening sprint; after a quiet month, every reliability ticket is pushed back again. An SLO replaces both reactions with a number the business agreed in advance, which is cheaper than arguing about each incident.
Which SLOs should a business application start with?
Start with the journeys that lose money or trust when they break, and measure them where the user meets the system: at the load balancer, API gateway or application edge. Two indicators per journey are enough at first: availability (the share of requests that do not fail on the server side) and latency (the share that finish within a threshold). An OpenTelemetry histogram of request durations gives both.
A starter set to use as a planning template, by application type:
-
Customer portal or storefront
- Indicator to start with
- Successful page and checkout requests; checkout latency under a threshold
- Where to measure
- Edge or load balancer logs
- Starting objective band
- 99.5% to 99.9% over four weeks
-
Internal ERP or back office
- Indicator to start with
- Successful requests during business hours; report generation within a threshold
- Where to measure
- Application request metrics
- Starting objective band
- 99% to 99.5% during working hours
-
API or integration endpoint
- Indicator to start with
- Successful calls; response time for the main operations
- Where to measure
- API gateway metrics
- Starting objective band
- Match what the consuming systems need, never above the weakest dependency
-
Batch and scheduled jobs
- Indicator to start with
- Jobs completed on time; records processed without error
- Where to measure
- Job scheduler results
- Starting objective band
- A share of runs finished before the business deadline
The bands are a starting point to argue about, not a recommendation for your service. Set the first objective close to what the service already achieves over the last few weeks, then tighten it only when users or the business need more. The Google SRE Workbook advises starting with an aspect that is relevant and easy to measure, and refining later.
Leave out indicators users never feel, such as CPU load or the health of a single server. They belong in dashboards for diagnosis, not in objectives.
How do you use an error budget in release planning?
The budget only works if product and engineering agree in advance what each state means for the next sprint or release. Write the rule as three states and keep it on one page:
| Budget state | Product owner | Engineering | Exception granted by |
|---|---|---|---|
| Healthy: most of the window's budget left | Plans features at the normal pace | Ships on the normal release cadence; may try riskier changes | Not needed |
| Burning: spend rate would exhaust the budget before the window ends | Keeps the next release small; accepts one reliability item in the sprint | Reviews the cause of the burn; adds a fix or a safer rollout to the plan | Head of engineering |
| Exhausted: the budget for the window is gone | Pauses non-urgent features until the service is back within its objective | Ships only fixes, security patches and reliability work | Named executive (CTO or CEO) |
Google's example error budget policy halts all changes and releases other than urgent fixes and security patches when the service has exceeded its budget over the preceding four weeks, and escalates disagreements about the calculation to the CTO. A small business can soften the first rule; it should keep the second, because a policy with no named decider is ignored under deadline pressure.
Review the budget at a fixed meeting. A weekly delivery review is enough: one line per SLO with the remaining budget, what spent it, and whether the next release changes plan.
Setting up SLOs with a small team
-
Pick the journeys
List the user journeys and pick the two or three that cost revenue or trust when they fail.
-
Define the indicators
Write each SLI as good events divided by valid events, measured at the edge.
-
Read the history
Compute the last four weeks from existing logs or metrics before choosing a target.
-
Set the objective
Choose a four-week rolling window and a target close to current performance.
-
Agree the policy
Write the three-state budget rule and get the product owner and a named executive to sign it.
-
Alert on burn rate
Page only when the budget is burning fast; open a ticket when it burns slowly.
-
Review and adjust
Read the budget at each weekly review and revisit targets once a quarter.
The SRE Workbook gives reasonable starting thresholds for a 99.9% objective: page when 2% of the budget is spent in one hour or 5% in six hours, and open a ticket when 10% is spent in three days. Alerting on budget burn instead of on every error spike is what makes SLOs workable for a team that has no dedicated on-call staff.
When should you not set SLOs yet?
SLOs are the wrong first step for a product without real traffic, because a few requests make every percentage noisy. They also fail when no one owns the service, when the team cannot measure request success at the edge, or when the application is an internal tool used by a handful of people who will simply call when it breaks.
Do not use an SLO as a promise to customers either. An SLA with penalties should sit below the internal objective, so the team sees the budget burning before the contract is at risk. If the business needs a contractual figure first, set the internal SLO, measure it for a quarter, and only then agree the SLA.
Where AI fits in SLO work
AI helps with the reading, not the deciding. An AI assistant can summarise what spent the error budget in a window by grouping incidents, deploys and alerts, and draft the weekly budget line for the delivery review, so the meeting starts from a prepared timeline. Netbase works with the major commercial and open-source models and chooses them per project.
Engineers still define every indicator, set the objectives with the product owner and confirm each cause in the metrics and logs themselves. A model never changes an alert rule or grants a budget exception.
- In delivered work
MLOps pipeline
An MLOps pipeline Netbase delivered for a client that is not named; an AI delivery record, not an SLO rollout.
- Available capability
AI agents in operations
AI-drafted error budget reviews are a way Netbase approaches the work, not yet tied to a published case.
What delivery record exists, and what does not
Netbase works as offshore development partner on a multi-tenant cloud ERP delivered as SaaS to small and mid-sized businesses for a US client, not named, and builds and runs its own products, such as Cloodo Workspace. Netbase names AWS, Google Cloud, DigitalOcean and Cloudflare as the cloud providers it uses, lists disaster recovery among its security practices, and offers SLAs on request.
Delivery runs in a six-step lifecycle with weekly reviews and KPI dashboards, which is where an error budget line belongs. Most Netbase projects are delivered on fixed-price contracts. No published Netbase record describes an SLO programme for a client, and no uptime, incident count or recovery time is claimed.
Alternatives: other ways to manage reliability
| Alternative | Fits when | Selection criterion |
|---|---|---|
| Uptime checks only | One service, low traffic, no release pressure | You only need to know that the site is down |
| Contractual SLA without internal SLOs | A customer demands a figure before you can measure | Accept that you will learn about breaches late |
| Starter SLOs with a budget policy | A few revenue journeys, a small team, regular releases | You want a planning rule without new headcount |
| A dedicated SRE function | Many services, strict contracts, constant on-call load | The cost of outages justifies a permanent team |
Teams that want to see where SLOs sit among their other practices can score them with the DevOps maturity assessment scorecard, and teams that still lack the request metrics an SLI needs can start with observability stack design with OpenTelemetry. When releases, pipelines and environments need work as well, Netbase's DevOps and platform engineering service covers the setup, and cloud migration and modernization is the place to define objectives before a move.
What are the limits of this guide?
- The objective bands are planning starting points, not benchmarks; your own history decides the target.
- The budget policy is a template; contracts, regulators or customers may require stricter rules.
- Burn-rate thresholds from the SRE Workbook assume steady traffic; low-traffic services need longer windows.
Plan the next step with a Netbase consultant
Frequently asked questions
No. A product owner, an engineering lead and existing request metrics are enough to start. The work is writing a few objectives, showing them on a dashboard and agreeing one error budget rule that both sides follow.
Start with two or three user journeys and two indicators each, usually availability and latency. More objectives than the team reads at its weekly review are noise and will be ignored.
The team pauses non-urgent feature releases and spends the next sprint on fixes and reliability work until the service is back within its objective. A named executive can grant an exception for a release the business cannot delay.
No. Set the internal objective tighter than the contractual SLA, so the team sees the budget burning and can react before a customer commitment is breached.
Next step
If you want a second opinion on your first objectives or your error budget policy, book a solution review with Netbase.
Related services and solutions
AI-ready cloud platform engineering for SaaS and commerce workloads Netbase provides cloud platform engineering for SaaS and commerce teams that need a secure, repeatable home for their software and its AI features, on AWS, Google Cloud, DigitalOcean or Cloudflare. We design the architecture, write the environments as code, connect managed and AI services and hand over a platform your team can run, built the way we run ours.
Learn More
AI-assisted cloud migration with a cut-over plan and a way back Netbase provides cloud migration services for companies running business applications on ageing servers or hosting to move them to AWS, Google Cloud, DigitalOcean or Cloudflare without losing data or control. AI speeds up the code and dependency assessment; every migration still has an inventory, a rehearsed cut-over, cost controls from day one and a tested rollback point.
Learn More
AI-assisted DevOps and platform engineering for routine releases Netbase provides DevOps and platform engineering for SaaS and commerce teams to make releases routine instead of risky. We set up CI/CD pipelines, repeatable environments and observability on AWS, Google Cloud, DigitalOcean or Cloudflare, with AI assistance for tests, reviews and incident triage, so every change ships the same way and problems surface before customers notice.
Learn More
SaaS product accelerator: launch an AI-ready SaaS on proven Netbase modules A SaaS product accelerator is a set of reusable Netbase modules for accounts, billing, roles and integrations that helps founders and product teams launch subscription software faster, with room for in-product AI from the first release. Reusing these modules can shorten development time, and the approach is proven on Printcart, Netbase's own web-to-print SaaS.
Learn More
Work with Netbase remotely
Every Netbase engagement is delivered remotely, so clients in any country and any industry can work with us; Netbase has worked with clients in 80 countries. Contracts and invoices are issued by Netbase JSC, and payment is made by bank transfer to Netbase JSC's company account or PayPal Business. Talk to the CEO, Huy Nguyen (David), on WhatsApp or by email.
Discuss a project
Netbase JSC helps organizations design, build, modernize, and operate digital products and AI-enabled business systems.project@netbasejsc.com
+84 937 869 689
91 Nguyen Chi Thanh, Dong Da, Hanoi, Vietnam
Get in touch
Tell us what you want to build, modernize, or operate.