Trust
Service levels
The response times and availability targets on our managed engagements, with the credits that apply when we miss. Published because a service level you have to ask for is not a commitment.
Last updated 1 July 2026 · Exinary Technologies Private Limited
Response times by severity
Response time means a qualified engineer is engaged and working — not an automated acknowledgement. Severity is assigned on impact, and you can escalate our assessment.
| Severity | Definition | NoOps | Managed DevOps |
|---|---|---|---|
| P1 — Critical | Production down, or a security incident in progress. Business operations stopped. | 15 minutes, 24x7 | 1 business hour |
| P2 — High | Major function degraded or a significant subset of users affected. No workaround. | 1 hour, 24x7 | 4 business hours |
| P3 — Medium | Partial degradation with a workaround available, or a non-urgent defect. | 4 business hours | 1 business day |
| P4 — Low | Question, request, or cosmetic issue. | 1 business day | 2 business days |
Business hours are 09:30–18:30 IST, Monday to Friday, excluding public holidays in Karnataka. NoOps P1 and P2 coverage is genuinely 24x7 including holidays.
Resolution targets
- P1: Continuous work until resolved or a workaround is live
- P2: Resolution or workaround within 8 hours
- P3: Resolution within 5 business days
- P4: Scheduled into the next work cycle
Resolution targets are targets, not guarantees, because some causes sit outside our control — a cloud provider outage, an upstream third-party failure, or a defect in your application code that needs your engineers. We commit to continuous progress and honest communication in those cases, and we will say plainly when the ball is not in our court.
Availability
The standard NoOps availability target is 99.95% monthly for the platform we operate, measured on end-user-facing availability rather than on individual component uptime. Higher targets are available where the architecture supports them, agreed per engagement.
Excluded from the calculation:
- Scheduled maintenance in windows agreed with you in advance, capped at 4 hours per month
- Failures in your application code, where our platform is operating correctly
- Force majeure and cloud provider regional outages beyond our designed redundancy
- Traffic exceeding capacity you declined to provision after we recommended it in writing
We will not sign an availability target on an estate we have not assessed. Every managed engagement starts with a two-to-three week assessment, and if the architecture cannot hold the number we will say so and quote the remediation separately.
Service credits
| Monthly availability | Credit against the next invoice |
|---|---|
| 99.95% or above | None — target met |
| 99.00% to 99.94% | 10% |
| 95.00% to 98.99% | 25% |
| Below 95.00% | 50% |
Repeatedly missed P1 response times attract a 10% credit independently of availability.
Credits are applied automatically — you do not have to claim them. We report attainment monthly with the raw incident data so you can verify the figure rather than take our word for it. See refunds and cancellation.
What is in scope
In scope: everything from the commit to the customer — pipelines, infrastructure, runtime, monitoring, scaling, patching and incident response — plus AI systems we built and operate, including their evaluation and cost monitoring.
Out of scope: your application code. If the fault is in your business logic we will diagnose it, page the right person and help — but we will not change it silently. Third-party SaaS you own, and anything a provider has broken upstream, are also outside our control though we will manage the escalation.
Escalation
Every managed engagement gets a named engagement lead and a defined escalation path with real names, not a shared inbox. Raise incidents via the agreed channel; for P1 use the 24x7 route given at onboarding rather than email.
If you are not getting the response you expect, escalate to the engagement lead and then to the department head. Contact details are in your onboarding pack, and [email protected] is monitored 24x7.
Questions?
Email [email protected] for anything about data or these policies, or [email protected] for security matters. See also our security page.