★★★★★4.7Google ratingPan-IndiaClients across IndiaCS-LedCA · CS · Advocate networkJanakpuriNew Delhi officeTransparentPricing, no hidden fees
ComplyKart Services
Service Level Agreement
Promises with teeth. SLAs that hold vendors accountable.
Every services contract contains a promise about quality. Most of them phrase it as an aspiration: "the provider shall use reasonable efforts to maintain high availability." That sentence has never kept a server online. A Service Level Agreement replaces aspiration with arithmetic — 99.9% monthly availability, P1 response within 30 minutes, resolution within 4 hours — and attaches a price to failure: service credits, escalation, and ultimately the right to walk away.
The SLA is where the commer
CA/CS-guided reviewYour documents and structure are checked before anything is filed.
Written scope firstInclusions and payable charges are explained before work begins.
Support till it’s doneOne team owns your filing, from the first call to the final handover.
Price on request
This work is scoped case by case. You get a fixed written quote before any work starts — no charge for asking.
Five steps, no guesswork — you always know exactly what comes next. Tap a step to see what it involves.
12345
You pay securely. Payment is confirmed instantly, and your receipt with the written fee breakup lands in your inbox.
Is an SLA the right instrument for you?
You need an SLA when:
You are buying any service where downtime or delay costs you money — cloud infrastructure, SaaS, managed IT, payment processing, logistics, customer support outsourcing.
You are a provider who wants to define quality commitments precisely rather than argue about "reasonable efforts" after every incident.
You operate in a regulated sector (banking, insurance, healthcare) where your regulator expects vendor SLAs — the RBI's IT outsourcing guidelines effectively require them for banks.
The contract value or the business criticality justifies the negotiation effort — which is a lower bar than most people think.
An SLA alone is not enough when:
There is no master agreement behind it — the SLA needs the MSA's framework (liability, term, termination, governing law) to be enforceable as part of a coherent contract.
The "service levels" are subjective — "high quality support," "best efforts" — with no numbers. That is not an SLA; it is marketing copy inside a contract.
Nobody will monitor compliance. An SLA with fifteen metrics and no one reading the monthly reports is theatre. Draft the SLA you will actually police.
What must be decided before drafting
Which metrics actually matter. Not every measurable thing deserves a service level. For infrastructure: availability and incident response. For support: response and resolution by severity. For professional services: milestone delivery and acceptance. Pick the three to five metrics whose failure would genuinely hurt, and make those the SLA. Everything else is reporting.
The numbers. What uptime percentage, what response times, what resolution times. These should reflect business impact and market reality — 99.9% for standard business services, higher for critical infrastructure — not the vendor's marketing sheet.
Who measures, and how disputes are settled. The provider's monitoring says 99.95%; your monitoring says 99.7%. Whose number governs? The SLA must name the measurement source, the methodology, and a dispute process — otherwise every credit claim becomes an argument.
The remedy structure. Service credits: the tiers, the percentages, the cap, whether credits are the exclusive remedy. And the chronic-failure trigger: how many bad months give you the right to terminate.
Exclusions you will accept. Scheduled maintenance, force majeure, customer-caused incidents — some exclusions are legitimate. Decide in advance which ones you will accept and what notice and boundaries each requires.
Information the drafter needs: the services in scope, the business impact of downtime or delay, the vendor's standard SLA (to mark up), your monitoring capabilities, regulatory requirements (if any), and your walk-away position on credits and termination.
Advantages of a properly drafted SLA
It converts "trust us" into arithmetic
"High availability" means nothing until it is 99.9% measured monthly, excluding only scheduled maintenance with 72 hours' notice. Numbers remove the argument about whether the service was good enough — the meter ran, the number is the number.
Service credits create the incentive structure
A provider's behavior follows the contract's economics. Credits that start at 10% of monthly fees and scale to 50% make outages expensive for the vendor — which is exactly when the vendor invests in redundancy. Credits capped at 5% make outages a rounding error.
It defines what "urgent" means before the crisis
At 3 a.m. during an outage, nobody should be debating whether this is a P1 or a P2. The severity matrix — P1 through P4, with response and resolution targets for each — is agreed in calm weather and applied in storms.
Measurement discipline improves the service itself
Providers who must report monthly against defined metrics run better operations. The reporting obligation — availability figures, incident counts, response-time compliance — creates an operational rhythm that benefits both sides beyond the credits.
Chronic-failure termination gives you an exit that credits cannot
Service credits compensate; they do not fix a structurally unreliable service. The termination right for repeated misses — three bad months in a rolling six, for example — is the clause that lets you leave a failing relationship instead of collecting coupons from it.
Regulators and auditors accept it as evidence
For regulated clients, a documented SLA with monitoring and reporting demonstrates vendor governance to auditors and regulators. The RBI's outsourcing guidelines, in particular, expect banks to hold their IT vendors to defined service levels.
It shortens incident disputes from weeks to hours
When the SLA defines severity, response targets, escalation paths, and communication duties, the post-incident argument collapses to a short factual check: what was the severity, when was it acknowledged, when was it resolved. That is a one-hour conversation, not a three-week dispute.
The anatomy of an SLA — clause by clause
Scope and definitions. Which services the SLA covers — all services under the MSA, or specific work orders. Define "availability," "downtime," "incident," "response," "resolution" precisely. The entire SLA stands on these definitions; ambiguity here infects every metric.
Service levels: availability. The uptime commitment, expressed as a monthly percentage. State the measurement period (calendar month), the formula (total minutes minus downtime, divided by total minutes), and exactly what counts as downtime — unavailability of the service to perform its core functions, as measured by the agreed monitoring.
Service levels: responsiveness. The severity matrix:
*P1 — Critical:* service unusable or major feature down. Response: 30 minutes. Resolution: 4 hours.
*P3 — Medium:* minor issue, workaround exists. Response: 8 hours. Resolution: 2–3 business days.
*P4 — Low:* cosmetic, informational. Response: 1–2 business days. Resolution: next release or as scheduled.
Adjust the numbers to the service; the structure is what matters.
Measurement and reporting. Who measures (provider tooling, customer tooling, or third party), the methodology, and the monthly report's contents and delivery date (within 10 days of month-end is standard). Include the customer's audit right over measurement data.
Service credits. The heart of the remedy structure:
*Trigger tiers:* e.g., availability below 99.9% but at or above 99.0% → 10% credit; below 99.0% but at or above 95.0% → 25% credit; below 95.0% → 50% credit.
*Response-time misses:* credit per P1 breach, or aggregated monthly.
*Cap:* commonly 30–50% of monthly fees per month.
*Application:* against future invoices, within a defined period; unapplied credits on termination convert to a final invoice credit or refund.
*Exclusivity:* state whether credits are the sole remedy for SLA misses or whether damages remain available. Vendors will push for exclusivity; resist it for wilful or repeated failures.
Exclusions. Scheduled maintenance (with minimum notice — 72 hours standard — and monthly hour caps), force majeure, customer-caused incidents, third-party failures beyond reasonable control, beta services. Each exclusion should be bounded: maintenance windows cannot be used to hide chronic unreliability.
Escalation. Named escalation contacts on both sides, with time-bound escalation triggers (e.g., unresolved P1 after 2 hours escalates to VP level). This clause gets used more than any other in the SLA.
Chronic failure and termination. The exit right: e.g., failure to meet service levels in any three months of a rolling six-month period gives the customer the right to terminate the affected services (or the agreement) on 30 days' notice, with transition assistance. Define whether termination is for the affected service only or the whole agreement.
Review. Service levels should be reviewed periodically — annually is standard — and adjusted as the service evolves. An SLA frozen for five years will be irrelevant by year three.
SLA vs MSA — the distinction that matters
Prospects search for both, and vendors sometimes blur them deliberately. Here is the clean distinction:
The MSA (Master Services Agreement) is the umbrella. It contains everything that makes the relationship a contract: the scope of services, the fees and payment terms, the term and renewal, intellectual property ownership, confidentiality, the liability cap and indemnities, termination rights, and governing law. It is negotiated once and changes rarely. Without it, you have no contract — just a quality promise floating in air.
The SLA is the quality annexure. It contains only the measurable promises and their remedies: uptime percentages, response and resolution matrices, measurement methodology, service credits, exclusions, escalation. It is negotiated per service or service tier and reviewed regularly. Without it, you have a contract with no measurable quality standard — just "reasonable efforts."
How they connect: the MSA should state that the SLA forms part of the agreement, that service credits are (or are not) the exclusive remedy for SLA misses, and that chronic SLA failure triggers the MSA's termination provisions. The two documents must be consistent on liability caps — a liability cap in the MSA that is lower than the SLA's credit cap creates a drafting contradiction a dispute will find.
The vendor's standard trick: presenting a "services agreement" that is really just an MSA with a decorative SLA — headline uptime, no credits, broad exclusions. When you see that, you are not looking at an SLA. You are looking at marketing.
Service credits — the mechanics that matter
This section deserves its depth because credits are where SLAs succeed or fail in practice.
Tiered, not flat. A single credit for "any miss" treats a 43-minute outage the same as a three-day one. Tier the credits to the severity of the miss: 10% for the first tier below target, 25% for deeper misses, 50% for catastrophic ones. The tiers should escalate steeply — the difference between 99.8% and 95% availability is the difference between an annoyance and a business crisis.
Calculated on the right base. Vendors often draft credits as a percentage of "the fees for the affected service" — which, for a multi-service contract, can shrink a 25% credit to a fraction of a percent of what you actually pay. Negotiate credits as a percentage of total monthly fees under the agreement, or at minimum the fees for the service line in scope.
Automatic, not claim-based. Many vendor SLAs require the customer to claim credits within 15–30 days of the incident, with supporting documentation — knowing most customers will not bother. Push for automatic credits: the provider's monthly report identifies the miss, the credit appears on the next invoice. At minimum, extend the claim window and simplify the process.
Caps that preserve bargaining power. A monthly cap of 30–50% of fees is standard and reasonable — it bounds the vendor's exposure while keeping misses painful. A cap of 5–10% is a vendor protection clause wearing a remedy's clothes. And watch the aggregate cap: some SLAs cap total credits over the contract term, which quietly deflates the remedy for long relationships.
Sole remedy — read it twice. Most vendor SLAs state that service credits are the customer's sole and exclusive remedy for SLA failures. Accept this for ordinary misses — it is market standard. But carve out wilful misconduct, fraud, data breaches, and repeated chronic failures, where damages and termination must remain available.
What credits do not do. Credits compensate; they do not fix. A provider can pay 50% credits every month indefinitely and still be a bad provider. That is why the chronic-failure termination right sits alongside the credits — the credits handle the bad month, the termination right handles the bad vendor.
Disputes we actually see
"Your monitoring is wrong." The provider's dashboard shows 99.95%; the customer's shows 99.7%. Without an agreed measurement source and methodology, every credit claim becomes a technical argument. Prevention: name the measurement tool and method in the SLA, and give the customer audit rights over the data.
The exclusion that ate the SLA. Scheduled maintenance defined so broadly — "any maintenance announced 24 hours in advance" — that the provider schedules around its outages. Or force majeure invoked for a foreseeable capacity failure. Prevention: bound every exclusion with notice periods, duration caps, and frequency limits.
Credits claimed, credits denied. The customer claims; the vendor disputes the severity classification or invokes an exclusion. Prevention: automatic credits based on the provider's own monthly report, with a short dispute window and a defined escalation path.
The sole-remedy trap. A major outage causes real business loss; the vendor points to the SLA's "sole remedy" clause and offers a 10% credit. Prevention: carve wilful misconduct, gross negligence, and data incidents out of the sole-remedy clause during negotiation.
Chronic failure without an exit. Eighteen months of missed service levels, thousands in accumulated credits, and no termination right — the customer is locked into a failing relationship. Prevention: the chronic-failure termination trigger, negotiated when you hold the bargaining power (before signing), not after.
Stamp duty and execution
Stamp duty on service agreements is a state subject. A standalone SLA — or an MSA with an SLA schedule — is typically stamped as a general agreement: ₹100 to ₹500 in most states.
Where the MSA involves substantial consideration and ongoing obligations, some states may assess duty differently — take advice on the specific state's position for high-value contracts. Execution: authorized signatories, designation and date, duplicate originals. Registration is not required.
The drafting process — step by step
Step 1 — The briefing call. We map the services, the business impact of failure, your monitoring capabilities, the vendor's standard terms, and your regulatory position. We identify the three to five metrics that actually matter.
Step 2 — SLA architecture. We design the metric set, the tiers, the credit structure, the exclusions you will accept, and the chronic-failure trigger. You approve the architecture before drafting.
Step 3 — First draft (or mark-up). For your own vendor contracts, we draft the SLA schedule. For vendor-provided SLAs, we mark up their standard terms — which is where most of our SLA work lives, and where the exclusions and credit mechanics need the closest reading.
Step 4 — Negotiation. SLA negotiation is technical and iterative — typically two to three rounds. We focus on the credit base, the exclusions, automatic vs. claim-based credits, and the termination trigger, because that is where the money and the bargaining power are.
Step 5 — Execution and operational handover. Final clean copy, stamping, signatures — and a handover note for your operations team: what to monitor, when reports are due, how to claim (if claim-based), and when the chronic-failure clock starts.
How long does it take?
A standalone SLA schedule for a standard services deal: 5–7 working days from briefing to execution-ready draft. Mark-up and negotiation of a vendor's standard SLA: 2–4 weeks depending on rounds. Complex multi-service or regulated-sector SLAs (banking, healthcare): allow longer, driven by the compliance review rather than the drafting.
Government fees and our fee structure
No government filing fee — the SLA is a private contract document. The only statutory cost is stamp duty (₹100–₹500 typically).
Our fee depends on the engagement: drafting an SLA schedule from scratch, marking up a vendor's standard SLA, or negotiating a full MSA-plus-SLA package. Fixed fee quoted after the briefing call.
How ComplyKart helps
Metrics that match the business. We start from your business impact — what downtime actually costs you — and design service levels around it, not around the vendor's template.
Credit mechanics with teeth. Tiered credits on the right fee base, automatic application, caps that preserve bargaining power, and sole-remedy carve-outs for the failures that matter. This is the part of the SLA vendors draft most carefully against you; we draft it back.
Exclusion surgery. We read every exclusion as a potential escape hatch and bound it — notice periods, duration caps, frequency limits — so the uptime promise survives its own fine print.
The exit right. Chronic-failure termination triggers negotiated when you hold the bargaining power, with transition assistance so you can actually leave.
Vendor-paper mark-ups. Most of our SLA work is improving the other side's standard terms. We know where the bodies are buried in vendor SLAs — the credit base, the claim windows, the exclusion drafting — and we go straight there.
What we will not do: we will not bless a vendor SLA with decorative metrics and no credits as "good enough" — if the remedy does not bite, we will tell you. We will not draft fifteen metrics nobody will monitor; we will argue for the five that matter. And we cannot manufacture bargaining power you do not have — against a hyperscaler with take-it-or-leave-it terms, we will tell you honestly what is negotiable and what is not, and focus your effort where it counts.
What happens after signing
In the first 30 days: set up monitoring against the SLA's metrics, confirm the monthly report format and delivery date with the provider, and brief your operations team on severity classifications and escalation contacts. The first month's report is the baseline — read it carefully.
Ongoing: review the monthly report against the SLA, track credit accrual, and watch the chronic-failure clock. Calendar the annual SLA review. Most SLA value is captured (or lost) in this operational discipline, not in the negotiation.
At renewal: renegotiate the service levels based on actual performance data. A provider who consistently beat 99.9% can be held to 99.95%; a provider who struggled at 99.9% should not be rewarded with looser terms.
Common mistakes and how to avoid them
Headline number, no credits. The classic vendor SLA: "99.9% uptime" with no service credits, or credits so small they do not matter. An SLA without meaningful credits is a press release. Always pair the promise with the remedy.
Credits on the wrong base. "10% of the affected service's fees" in a contract where the affected service is 2% of the spend. Negotiate the base as total monthly fees or the full service line.
Claim-based credits with short windows. "Customer must claim within 15 days with supporting evidence." Most misses go unclaimed. Push for automatic credits; at minimum, a 60–90 day claim window with a simple process.
Unbounded exclusions. Scheduled maintenance with 24 hours' notice and no monthly cap; force majeure covering foreseeable events. Bound every exclusion.
No chronic-failure exit. Credits without a termination right lock you into a failing vendor. The exit trigger is as important as the credit table.
Metrics nobody monitors. Fifteen service levels, no reporting discipline, no one reading the reports. Draft the SLA your team will actually operate — five metrics with monthly reporting beats fifteen with none.
Sole remedy accepted blindly. "Service credits are the sole remedy" — accepted without carve-outs for wilful misconduct, data breach, or gross negligence. Read it twice; carve out what matters.
Sources
This page is built on the Indian Contract Act, 1872 (the general law of contract governing service agreements), the Information Technology Act, 2000 and the Information Technology (Reasonable Security Practices and Procedures and Sensitive Personal Data or Information) Rules, 2011 (for data-handling aspects), RBI guidelines on outsourcing of IT services by banks and financial institutions, and current market practice for cloud and SaaS SLAs (uptime tiers, credit structures, and exclusion drafting as reflected in major providers' published terms and practitioner guides). Commercial terms described are market practice, not statutory requirements — actual numbers are negotiated per deal.
FAQ
Frequently asked questions
22 answers from our consultants.
You pay for our expert's time — document checking, filing, follow-ups — not the government. The government fee goes to the government; our fee is for the work we do so you don't have to figure it out yourself.
Tell us your situation in plain words — what you're trying to do, not which form you think you need. Half our job is stopping clients from buying the wrong service. The first consultation is free and honest.
Yes. We quote professional fee + government charges + taxes separately, in writing, before you pay. The number doesn't change midway. If the government fee changes, we show you the notification.
Every engagement gets a tracker — you see what's filed, what's pending and what's next. No chasing on WhatsApp for status updates.
Both. Startups get incorporation + early compliance bundles; established companies get annual retainers. About half our clients started with us at incorporation.
When the provider misses a service level, you get a credit — a percentage of the monthly fee applied against future invoices. Typical structure: 10% credit for the first miss tier, 25% for deeper misses, up to a monthly cap of 30–50%. Credits are usually the exclusive remedy for SLA misses, which is why their size and base matter so much.
99.9% monthly (about 43 minutes of permitted downtime) is the standard for business services. 99.95% (about 22 minutes) for enterprise tiers; 99.99% (about 4 minutes) for premium infrastructure. Below 99.5% (3.6 hours/month) is generally unacceptable for production workloads. Match the number to business impact, not marketing.
Scheduled maintenance (with notice requirements and duration caps), force majeure, customer-caused incidents, third-party failures outside the provider's control, and beta services. Every exclusion narrows the promise — bound each one with notice periods and caps so exclusions cannot swallow the commitment.
No. Credits reduce future invoices; they are not cash compensation for loss. Most SLAs cap credits and declare them the sole remedy for SLA misses. For losses beyond the credit — especially from wilful misconduct or data incidents — you need the sole-remedy carve-outs negotiated in advance.
The SLA must say. Options: the provider's monitoring, the customer's, or a third party. The provider's numbers govern in most vendor SLAs — which is why you need audit rights over the measurement data and a dispute process for conflicting readings. "Trust our dashboard" is not a measurement methodology.
The classification system for incidents: P1 (critical — service unusable), P2 (high — significant degradation), P3 (medium — minor, workaround exists), P4 (low — cosmetic). Each tier carries response and resolution targets — e.g., P1 response in 30 minutes, resolution in 4 hours. Agreed in calm weather, applied in storms.
Automatic. Claim-based credits with 15–30 day windows go unclaimed — vendors know this. Push for credits applied automatically based on the provider's own monthly report. If the vendor insists on claims, negotiate a 60–90 day window and a simple process.
Start at 5–10% of monthly fees for the first miss tier, scaling to 25–50% for severe outages, with a monthly cap of 30–50%. Below that range, the credits do not change vendor behavior. And check the base — the percentage must apply to total monthly fees or the full service line, not a sliver of it.
If the service matters to your operations, yes — scaled to size. A short SLA with one uptime promise, one response-time promise, and a simple credit table is enough. The question is not contract size but business impact: what does an outage cost you?
Replace uptime with deliverable-based metrics: milestone completion dates, acceptance criteria, and revision turnaround times. Credits become fee reductions or rework obligations. The structure — measurable promise, measurement, remedy, exclusions — is identical; only the metrics change.
The SLA should say. Best practice: unapplied credits convert to a credit against the final invoice, or are refunded if they exceed it. Without this clause, credits evaporate on termination — which perversely rewards the vendor for failing near the end of the term.
Standard vendor terms often reserve this right with notice. Resist it for material changes — service levels, credit percentages, and exclusions should require mutual agreement. At minimum, require 30–90 days' notice and give yourself a termination right if a change is adverse.
A concept from site-reliability practice: the permitted unreliability implied by the uptime target (e.g., 0.1% at 99.9%). Some SLAs let the provider "spend" the error budget on risky deployments. If your SLA uses this concept, define who decides how the budget is spent and what happens when it is exhausted early in the period.
For regulated clients, the SLA or MSA should include data-residency commitments (e.g., data stays in Indian regions), compliance with the IT Act's Reasonable Security Practices Rules, and — for banks — alignment with RBI outsourcing guidelines. These are contractual commitments, not just policies; breach should trigger the SLA's remedies.
Availability figures per service, incident counts by severity, response-time and resolution-time compliance percentages, scheduled vs. unscheduled downtime, credits accrued, and any exclusion invocations with justification. Delivered within 10 days of month-end. If the report does not contain what you need to verify compliance, the reporting clause needs rewriting.
Only if your business loses money in the gap between 99.9% (43 min/month) and 99.99% (4 min/month). For most business applications, 99.9% with strong response times and real credits beats 99.99% with weak remedies. Buy the remedy structure first, the extra nine second.
At renewal, yes — that is where your bargaining power lies. Come with performance data: the outages, the impact, the missing remedies. Mid-term, you can propose an amendment, but the vendor has little incentive to agree. Prevention remains the strategy: never sign an SLA without credits in the first place.
What clients say about working with us
Real stories from businesses we've worked with.
Trademark Registration“
Our Trademark was applied in all the classes with a short span of time by Complykart Trademark Experts and I can say the level of expertise and commitment they carry is remarkable. It's an A+ from my side. We are now sticking to ComplyKart only.
Company Registration“
I found the professionalism of the staff at complykart.com, a pleasure to work with. As a foreigner to India, they took the time to explain the documentation and assisted in every question I raised. I was pleased that complykart.com made it easy for my Business.
Merger & Acquisition“
Trust, Commitment, Dedication and responsiveness are the best things with ComplyKart. Thanks for handling our Merger assignment with so much care. After my business disputes, complykart handles all things with own sense of understanding.
Senior Consultant · ESOP, Trademark & General Advisory · 14+ years experience
Pankaj Jain is a senior consultant covering ESOP structuring, trademark and IP protection, accounting advisory and legal documentation. With the broadest practice at the firm, he handles the matters that don’t fit neatly in one box.
Last reviewed: 16 November 2026
Get a free consultation
Talk to a ComplyKart professional about Service Level Agreement. The first consultation is free.