Service Level Agreement
Last updated: August 22, 2026
This Service Level Agreement (the "SLA") sets out the uptime commitments, support response times and service credits Ankra AB provides for the Ankra Service. It forms part of the Agreement described in ourTerms of Service, and capitalised terms used here have the meaning given there.
You do not need to sign anything. The service levels for a paid plan apply automatically from the start of the Billing Period in which that plan takes effect, and stop applying if the plan is downgraded or the Agreement ends.
The Free plan carries no service levels. The Ankra Service is provided to Free plan users on an "as is" basis, as set out in clause 14 of the Terms of Service.
1. What each plan gets
1.1. Startup is our basic service level. Enterprise starts at a materially higher level and can be raised further by agreement. The table below is the whole commitment for Free and Startup customers.
| Free | Startup | Enterprise | |
|---|---|---|---|
| Applies | No service levels | Automatically, from the start of the paid plan | Automatically, from the start of the paid plan |
| Monthly uptime commitment | None - provided as-is | 99.5% | 99.9% |
| Service credits | None | Up to 25% of the month's Service Fees | Up to 50% of the month's Service Fees |
| Support channel | Community | Priority email | Named contact, email and a shared channel |
| Support hours | None | 09:00-17:00 CET, Monday to Friday | 09:00-17:00 CET, Monday to Friday; 24/7 available as a paid add-on |
| First response - Critical | None | 24 hours | 4 hours |
| First response - High | None | 24 hours | 12 hours |
| First response - Normal and Low | None | 2 business days | 1 business day |
| Included AI allowance, per month | USD 10 of Ankra-funded AI usage | USD 50 of Ankra-funded AI usage | USD 200 of Ankra-funded AI usage |
| AI beyond the allowance | Add your own model key | Add your own model key | Add your own model key, or a raised allowance by agreement |
| Playground environments | Subject to availability, no service levels | Subject to availability, no service levels | Subject to availability, no service levels |
| Incident communication | status.ankra.io | status.ankra.io | status.ankra.io, plus a written root-cause report within 5 business days of a Critical incident |
| Security incident notice | Within 72 hours of confirmation | Within 72 hours of confirmation | Within 48 hours of confirmation, with a written incident report within 5 business days |
| Scheduled maintenance | No commitment | 48 hours' notice | 48 hours' notice; at most 4 hours a month, outside 09:00-17:00 CET |
| Escalation | None | Reply on the open ticket | Named account contact, acknowledged within 2 business hours |
| Terms negotiable | No | No - published terms only | Yes - this page is the starting point, not the ceiling |
1.2. Enterprise is the only plan whose terms are negotiable. Where Ankra and an Enterprise Customer sign a separate written agreement, that agreement prevails over this SLA to the extent of any conflict. Free and Startup are supplied on these published terms without variation.
1.3. Enterprise Customers may additionally purchase 24/7 support coverage and on-premises deployment as paid add-ons. Those add-ons replace the support hours and response times in the table above with the ones stated in the relevant order.
2. Uptime
2.1. "Monthly Uptime" is the percentage of total minutes in a calendar month during which the core services of the Cloud Ankra Platform - the API, the dashboard, the GitOps reconciliation engine and cluster management - are available, calculated as:
Monthly Uptime = ((Total Minutes - Downtime Minutes) / Total Minutes) x 100
2.2. Downtime is measured from Ankra's monitoring atstatus.ankra.io, which is the reference record for this SLA. Downtime begins when an outage is first detected by that monitoring or reported by the Customer, whichever is earlier, and ends when the affected service is restored.
2.3. Workloads running inside the Customer's own clusters are not part of the Cloud Ankra Platform. A cluster continues to run its workloads when the Ankra control plane is unavailable, and the availability of the Customer's clusters, cloud accounts and applications is not measured by this SLA.
2.4. The AI Features are not part of the measured services. They depend on third-party model providers, and Ankra does not control their availability, latency or capacity. Reaching an AI allowance or a daily spend limit is not Downtime.
2.5. Playground environments are not part of the measured services. They are temporary evaluation and development environments provided subject to capacity, and they carry no uptime commitment on any plan.
3. Support severity and response
3.1. Ankra assigns a severity to each request on receipt, acting reasonably. Response times are first human response, not resolution.
| Severity | Meaning |
|---|---|
| Critical | The Ankra Platform is unavailable, or a defect in the Ankra Platform is preventing all deployments across the Customer's clusters, with no workaround. |
| High | A core function - deployments, GitOps reconciliation, cluster management or the API - is materially degraded or unavailable for one or more clusters, and no reasonable workaround exists. |
| Normal | A function behaves incorrectly or is degraded, and a reasonable workaround exists. |
| Low | Questions, configuration guidance, cosmetic defects, and feature requests. |
3.2. Response times run during the Customer's support hours as set out in clause 1. A request received outside those hours is treated as received at the start of the next support hours.
3.3. Support covers the Ankra Service itself. Diagnosing and fixing the Customer's own applications, infrastructure and cloud accounts is outside this SLA and is available separately as consultancy.
3.4. For Customers on the Enterprise plan, Ankra will provide a written root-cause report within 5 business days after resolving a Critical incident, and will notify a security incident affecting Customer Data within 48 hours of confirming it, followed by a written incident report within 5 business days. For all other plans, security incidents are notified within 72 hours of confirmation as set out in clause 18 of the Terms of Service and theData Processing Agreement.
4. Service credits
4.1. If Monthly Uptime falls below the commitment for the Customer's plan, the Customer may claim a service credit calculated as a percentage of the Service Fees for the affected month.
| Startup | |
|---|---|
| 99.5% or above | None |
| Below 99.5% but at or above 98.0% | 10% |
| Below 98.0% | 25% |
| Enterprise | |
|---|---|
| 99.9% or above | None |
| Below 99.9% but at or above 99.0% | 10% |
| Below 99.0% but at or above 95.0% | 25% |
| Below 95.0% | 50% |
4.2. To claim, email[email protected]within 30 days of the end of the affected month, identifying the organisation and the dates and times of the downtime. Ankra will confirm the claim against its monitoring record and apply any credit to the next invoice.
4.3. Credits are applied against future Service Fees. They are not refunds, are not exchangeable for cash, and cannot exceed the Service Fees payable for the affected month.
4.4. Service credits are the Customer's sole and exclusive remedy for a failure to meet the commitments in this SLA. Clause 16 of the Terms of Service continues to limit Ankra's liability in all other respects.
5. What is excluded
5.1. The following do not count as Downtime and do not give rise to service credits:
- scheduled maintenance announced at least 48 hours in advance (for Enterprise, limited to four hours in any calendar month and performed outside 09:00-17:00 CET), and emergency maintenance needed to address a security risk;
- failure, degradation or capacity limits of a cloud infrastructure provider, source code host, DNS provider, or AI model provider used by the Customer, including where the Customer's own account is suspended, throttled, out of quota or out of credit;
- anything caused by the Customer's own configuration, clusters, workloads, network, credentials, or use of the Ankra Service contrary to the Agreement or Ankra's documentation;
- the AI Features, including a model provider's own outage, latency, rate limiting or capacity shortfall, and the exhaustion of an AI allowance or daily spend limit;
- playground environments, including capacity refusals, expiry at the end of their lifetime, and deletion of their contents;
- features identified as alpha, beta, preview or experimental;
- suspension under clause 12 of the Terms of Service, or non-payment;
- force majeure, as described in clause 23.6 of the Terms of Service.
5.2. Because Ankra manages clusters that run on infrastructure the Customer owns, an incident affecting the Customer's cloud provider is not an Ankra outage even where it is visible through the Ankra Platform.
5.3. Playground environments are the exception to clause 5.2: they run on infrastructure Ankra operates and shares between customers. They are still excluded from this SLA, because they are evaluation and development environments rather than a production service. They are provided subject to capacity, are limited in number, expire automatically, and are deleted with their contents on expiry. Nothing in a playground environment is backed up, and Ankra cannot restore one after it has expired. Do not run production workloads or keep the only copy of anything there. The terms that govern them are in clause 6 of theTerms of Service.
6. Changes to this SLA
6.1. Ankra may update this SLA in line with clause 22 of the Terms of Service. Where a change materially reduces a commitment, it takes effect at the start of the Customer's next Renewal Term rather than immediately.
6.2. Where an Enterprise Customer has a signed agreement covering service levels, changes to this page do not alter that agreement.
7. Contact
For service level questions or credit claims, contact[email protected]. To discuss an Enterprise agreement or the 24/7 add-on, contact[email protected].