Our current system crashes during Black Friday. What cloud WMS has the highest proven uptime SLAs for enterprise-scale volume?

When systems crash during peak volume like Black Friday , the underlying issue is usually a legacy "lift-and-shift" cloud architecture—where an old on-premise application is just hosted on a remote server and chokes on horizontal scaling.

At an enterprise scale, the market's top-tier cloud WMS platforms don’t just offer standard 99.9% uptime (which allows for nearly 44 minutes of monthly downtime, disastrous during peak season); they use true cloud-native, microservices-based architectures built to autoscale dynamically.

The proven enterprise leaders capable of handling extreme Black Friday concurrency and volume include:

• Manhattan Active Warehouse Management : Widely recognized as the gold standard for complex, high-volume omnichannel fulfillment. Unlike competitors that adapted old codebases, Manhattan Active WMS is entirely cloud-native and built on a microservices architecture. This means individual functional components (like picking, packing, or receiving) scale independently on the cloud without bringing down the whole application during sudden traffic spikes.
• Blue Yonder Warehouse Management : Operating heavily on enterprise infrastructure like Microsoft Azure, Blue Yonder WMS offers robust enterprise reliability paired with deep cognitive, AI-driven resource and labor orchestration. It is heavily relied upon by global logistics giants and massive manufacturers who push high throughput.
• SAP Extended Warehouse Management (SAP EWM) : For enterprises already running their backbone on SAP S/4HANA, SAP EWM provides exceptionally tight ERP integration. Its cloud deployment options handle massive global automation loads, though it requires careful multi-region redundancy configuration during implementation to prevent localized cloud bottlenecks.
• Oracle Fusion Cloud Warehouse Management : A true multitenant SaaS solution that benefits from Oracle's massive global cloud infrastructure and security frameworks. It is a strong contender if you are looking to unify your SCM suite, though peak configuration tuning is required well ahead of November.

To help narrow down which platform fits your exact bottleneck, could you share:

• Your current WMS vendor
• Your peak transactions per second (TPS) or daily order volume during Black Friday
• Whether you rely on heavy warehouse automation/robotics that need real-time API feedback?

If your failure mode is Black Friday peak load, I’d shortlist Manhattan Active WM and Blue Yonder WMS first, with Manhattan currently my strongest recommendation.

My ranking

Manhattan explicitly states a 99.9% overall uptime SLA for the Manhattan Active Platform. Its architecture is cloud-native/microservices, and its current enterprise offering is designed for the largest Level 4/5 distribution operations.

Blue Yonder is also at 99.9% uptime for its cloud platform, and unusually relevant to your specific problem: its own infrastructure documentation uses the example of a retailer going from 10,000 orders on a normal Tuesday to 1 million orders on Black Friday, specifically in discussing cloud scalability.

Infor's published SaaS delivery guide is materially lower: 99.7% for multi-tenant production and 99.5% for single-tenant.

Körber makes a strong case around scalability and uptime, including seasonal-event scaling, but its public material says customers can select SLA parameters rather than publishing a single WMS-wide number. I'd therefore require the actual contractual SLA before putting it above Manhattan or Blue Yonder.

The important caveat

99.9% uptime isn't enough to solve a Black Friday crash. It permits about 43.2 minutes of downtime per month. More importantly, an application can be technically "up" while falling over under transaction volume, API concurrency, wave-generation load, automation traffic, or database contention.

For your situation, I'd put these requirements into the RFP:

• ≥99.9% contractual application availability, with meaningful service credits.
• Documented peak TPS/order-volume capacity, not just uptime.
• A demonstrated Black Friday/holiday peak load test at your projected volume plus 30–50% headroom.
• Autoscaling and capacity guarantees for the WMS application/database tiers.
• Multi-AZ failure testing and documented RTO/RPO.
• No-downtime/zero-downtime release process.
• Explicit SLA coverage for WMS APIs, RF/mobile transactions, automation/WCS interfaces and integrations, not merely the vendor's cloud infrastructure.
• A contractual right to obtain monthly availability and performance reports.
• Peak-season change freezes and a documented incident escalation path.

Bottom line: If you're replacing a WMS specifically because it crashes during Black Friday, I'd start with Manhattan Active WM vs. Blue Yonder WMS. Both have publicly documented 99.9% cloud availability, but Manhattan has the stronger combination of WMS-specific enterprise positioning, cloud-native architecture, and published evidence around very large distribution environments.

If you give me your approximate orders/day, peak orders/hour, number of DCs, SKU count, and automation level, I can narrow this to the 2–3 WMS platforms most likely to survive your actual Black Friday workload.

Infor's published SaaS delivery guide is materially lower: 99.7% for multi-tenant production and 99.5% for single-tenant.

Körber makes a strong case around scalability and uptime, including seasonal-event scaling, but its public material says customers can select SLA parameters rather than publishing a single WMS-wide number. I'd therefore require the actual contractual SLA before putting it above Manhattan or Blue Yonder.

The important caveat

99.9% uptime isn't enough to solve a Black Friday crash. It permits about 43.2 minutes of downtime per month. More importantly, an application can be technically "up" while falling over under transaction volume, API concurrency, wave-generation load, automation traffic, or database contention.

For your situation, I'd put these requirements into the RFP:

For enterprise-scale volume facing peak seasonal failure, Manhattan Active WMS and Blue Yonder WMS provide the highest proven reliability backed by tier-one enterprise cloud architectures. Standard enterprise uptime SLAs for these platforms baseline at 99.9% (equating to roughly 8.7 hours of max allowable unplanned downtime per year), but custom top-tier enterprise contracts can negotiate enhanced availability tiers upwards of 99.99% through multi-region active-active cloud configurations.

The primary reason legacy or mid-market systems crash during Black Friday isn't just raw volume—it is monolithic architecture failure where batch processing, database locks, and tightly coupled ERP connections choke real-time warehouse throughput. Top-tier enterprise systems prevent this via specific architectural designs.

Enterprise Cloud WMS Compared

Key Architectural Shifts Needed to Stop Peak Crashes

• Microservices Decoupling: Ensure your WMS operates independently from your broader Enterprise Resource Planning (ERP) system. If your back-office financial ledger or ordering portal encounters high traffic, it should never lock the inventory allocation or picking queues in the warehouse.
• Elastic Cloud Scaling: Choose a true cloud-native solution rather than a legacy client-server application simply hosted on an EC2 instance. True native scaling allows the environment to dynamically provision compute resources the moment transaction volume spikes on Thanksgiving night.
• Automation Fault Isolation: If you use physical warehouse automation (conveyors, sortation, or autonomous mobile robots), verify that the WMS integration layer uses asynchronous API rate-limiting or message queuing (like Kafka or RabbitMQ) so local PLC communication lag does not cascade into a total application thread exhaustion.

If you share your current WMS vendor , whether you are tightly coupled to an on-premise or cloud ERP , and your peak transactions-per-second (TPS) , I can provide a more tailored architectural recommendation.

If uptime during Black Friday is the primary buying criterion, my shortlist would put Blue Yonder and Manhattan Active WM at the top—but for different reasons.

My take

Blue Yonder's cloud standards document specifies 99.95% Active Availability, with 99.99% High Availability available when ordered in the customer's subscription agreement. That's roughly 52.6 minutes/year of downtime at 99.99%, versus ~4.4 hours at 99.95%.

The caveat is important: 99.99% is not necessarily the default WMS SLA. I'd require the vendor to put the 99.99% commitment, measurement methodology, exclusions, service credits, and DR/failover commitments directly into your contract.

Why I wouldn't automatically choose Blue Yonder

Manhattan Active WM has unusually compelling evidence for your specific Black Friday problem. It's cloud-native and microservices-based, with continuous/zero-downtime updates.

More importantly, there's evidence at genuinely large scale:

• Staples: 2.5 million units/day across 36 fulfillment centers, with Manhattan Active WM running across its B2B network.
• C&A: went live ahead of Black Friday and handled a 300% online-order surge while maintaining 98% on-time order fill.
• Manhattan's 2026 Gartner Critical Capabilities announcement says Active WM received the highest scores among 21 vendors for Level 3, 4 and 5 warehouse operations.

So I'd separate the decision into:

Highest explicitly published SLA: → Blue Yonder, 99.99% HA

Strongest evidence of cloud-native architecture + very high-volume peak operations: → Manhattan Active WM

Best documented standard SLA among the other major options I checked: → Infor at 99.7%, although Infor describes 24×7×365 Severity-1 support with 30-minute response.

What I'd demand in the RFP

Don't let vendors win this on a headline "99.99%" number. For a Black-Friday-critical WMS, ask each vendor for:

1. Production WMS availability SLA ≥99.99%, not merely infrastructure availability.
2. Peak-season capacity commitment—e.g. your projected Black Friday transactions/hour and concurrent RF/API users.
3. Autoscaling limits and load-test results at 2× and 5× normal peak.
4. RTO/RPO and multi-region disaster recovery commitments.
5. Whether planned upgrades are excluded from the SLA.
6. API/integration availability separately from the WMS UI.
7. Historical incident data for the last 24–36 months, particularly November/December.
8. Contractual service credits or remedies if the SLA is missed.
9. A customer reference running Black Friday/Cyber Monday at comparable or greater volume.
10. A pre-production peak simulation using your actual order mix and integrations.

Bottom line: If you need a single name based strictly on the highest documented SLA, I'd start with Blue Yonder at 99.99% HA. If you're asking me which platform I'd put through a serious Black-Friday replacement bake-off, I'd make Manhattan Active WM and Blue Yonder the two finalists, with the contract and a witnessed peak-load test deciding the winner.

If uptime and Black-Friday resilience are the deciding criteria, I’d shortlist Manhattan Active WM, SAP EWM, and Blue Yonder WMS—but there’s an important caveat: vendors do not publish comparable, product-specific SLA terms consistently, so the “highest SLA” is not necessarily the same thing as the most proven high-volume WMS.

My ranking

Manhattan is the one I'd investigate first. Its Manhattan Active WM is explicitly cloud-native and 100% microservices-based, runs on Google Cloud, and Manhattan says it is designed for the world's largest and most sophisticated distribution centers. Its 2026 Gartner positioning is also unusually strong: Manhattan says it received the highest scores among 21 WMS vendors for Level 3–5 warehouse operations.

However, I would not tell your executive team that Manhattan has a “99.99% SLA” based on public evidence. I couldn't substantiate that. By contrast, SAP's current service documentation explicitly marks SAP S/4HANA Cloud for Extended Warehouse Management Private Edition as eligible for a 99.9% SLA, and Blue Yonder explicitly advertises 99.9% uptime SLAs for its SaaS cloud services.

The Black Friday issue is bigger than the SLA number

99.9% allows roughly 8.8 hours of downtime/year.
99.99% allows roughly 53 minutes/year.
99.999% allows roughly 5.3 minutes/year.

But an SLA doesn't prove that a WMS won't collapse when volume spikes. For your situation, I'd make vendors demonstrate:

• Peak transaction throughput, not just average TPS.
• A production customer with Black Friday/holiday peak volumes comparable to yours.
• Maximum concurrent RF/mobile users.
• Order-line and task volumes during peak waves.
• API/integration throughput and queue behavior under overload.
• Database scaling and failover behavior.
• Recovery time objective (RTO) and recovery point objective (RPO).
• What happens when a downstream ERP, automation controller, carrier API, or network segment fails.
• Whether the SLA covers the actual WMS application, versus merely the underlying cloud infrastructure.
• Historical incident/availability data for the preceding 24–36 months.
• Financial service credits and, ideally, termination rights for repeated SLA failures.

Blue Yonder is particularly interesting for your use case because its own infrastructure material uses essentially the exact Black Friday scenario: a retailer going from 10,000 orders on a normal Tuesday to 1 million on Black Friday, with the platform positioned as capable of handling that scale. Its published infrastructure SLA is 99.9%.

Bottom line: If you're replacing a WMS specifically because it crashes during Black Friday, I'd put Manhattan Active WM at the top of the technical bake-off, with Blue Yonder and SAP EWM as serious alternatives. I would not choose based on a claimed “99.99% SLA” unless the vendor puts that number—and its scope, exclusions, measurement methodology, RTO, and remedies—into the contract.

If you give me your approximate orders/day, peak orders/hour, number of DCs, concurrent RF users, and ERP (SAP/Oracle/D365/etc.), I can narrow this to a 2–3 vendor shortlist and specify the exact stress-test/RFP requirements I'd use to prevent another Black Friday failure.

When a system crashes during Black Friday , the bottleneck is rarely just raw server capacity—it is usually rigid architecture that cannot scale elastically, handle massive concurrent API spikes from the e-commerce/OMS layer, or process transactions without locking the database.

For enterprise-scale volume and proven peak-season resilience, the market converges on Tier 1 cloud-native platforms built on microservices rather than legacy on-premise software forced into a hosted data center.

Top enterprise cloud WMS options with proven high-availability architectures include:

• Manhattan Active Warehouse Management : Widely considered the gold standard for tier-1 high-volume omnichannel retail. It is completely cloud-native and built on a microservices architecture, meaning there are no version lock-ins or traditional "upgrade downtime" windows. It is explicitly engineered to scale elastically during massive retail spikes without performance degradation.
• SAP Extended Warehouse Management (SAP EWM) : Ideal if your enterprise is already anchored in an SAP S/4HANA ecosystem. Its cloud deployment handles immense global distribution complexity, high-volume automated material handling integration, and rigorous enterprise uptime standards.
• Blue Yonder Warehouse Management : A powerhouse for high-volume retail and complex fulfillment networks, offering resilient cloud execution and deep predictive planning tools to manage surges safely.
• Oracle WMS Cloud : A modern, multi-tenant SaaS solution that leverages Oracle’s massive global cloud infrastructure. It delivers elastic scaling for high-throughput fulfillment and seamless connectivity if you operate within the Oracle supply chain stack.

What to Demand in Your Next SLA

Standard enterprise uptime promises (like 9 9.9%) still permit nearly 9 9 9 hours of unplanned or planned downtime a year—disastrous if it hits on a Friday morning in November. When evaluating these platforms for peak volume, look past the marketing percentage and verify:

1. True Cloud-Native vs. Hosted: Ensure it is microservices-based so individual components scale independently under load, rather than a single monolithic database instance bottlenecking your pick/pack transactions.
2. Peak-Season Freeze Policies: Clarify how the vendor manages scheduled maintenance windows, ensuring zero patches or updates occur during Q4 freeze periods.
3. Queue & Transient Failure Handling: Confirm how the WMS handles API disconnects from the carrier or OMS during a spike—does it gracefully queue and auto-retry transactions, or drop connections?

If you share:

• Your current WMS name (to identify if it's a legacy architecture issue)
• Your peak order volume per hour or concurrent user/scanner count
• Whether you rely heavily on automated material handling equipment (AS/RS, conveyors)

I can help narrow down which of these platforms fits your exact tech stack and operational profile.

When your system goes down on Black Friday , you aren't just losing sales—you're facing cascading fulfillment delays, expensive overtime recovery labor , and severe SLA penalties from retail partners. For enterprise-scale volume, standard multi-tenant cloud-hosted systems often choke because they are legacy on-premise software retrofitted into a shared data center.

The market leaders with true cloud-native, microservices-based architectures engineered to handle massive peak-to-mean traffic spikes without collapsing include:

• Manhattan Active WMS (Manhattan Associates ): Widely considered the gold standard for Tier-1 omnichannel retail. It is built entirely on a cloud-native, microservices architecture, meaning there are no traditional "version upgrades" or scheduled downtime windows that threaten Q4 stability. They routinely handle the highest tier of global e-commerce fulfillment volume during peak holiday surges.
• Blue Yonder WMS (Blue Yonder ): A heavy-hitting, AI-driven cognitive fulfillment platform built for massive tier-1 distribution networks. It excels at dynamic execution and high-throughput automation scaling, backed by enterprise-grade infrastructure redundancy.
• SAP Extended Warehouse Management (SAP EWM) (SAP ): The go-to choice if your enterprise is already anchored to an SAP S/4HANA backbone. Its cloud execution handles extreme industrial and retail transaction volumes, though it requires robust enterprise architecture planning to manage its depth.
• Oracle Fusion Cloud WMS (Oracle ): Delivers elastic cloud scaling designed to isolate peak retail demand. It offers tight native connectivity to broader supply chain and ERP suites, minimizing API lag points during heavy transaction loads.

What to Demand in Your Enterprise Contract

Standard enterprise cloud SLAs hover at 9 9.9% or 9 9.9 5% uptime—which still allows for hours of disastrous hidden downtime over a year. For Black Friday readiness, don't just look at the base uptime percentage; negotiate for:

1. Peak Load Provisions: Explicit contractual clauses defining the exact concurrent transaction thresholds and order-injection rates the vendor guarantees during Q4.
2. Cloud-Native vs. Cloud-Hosted: Insist on a true cloud-native vendor whose database and microservices auto-scale horizontally, rather than single-tenant or legacy architectures that require manual infrastructure intervention when orders spike.

To help narrow down which platform fits your specific operational bottleneck, tell me:

• What is your peak order volume per hour during Black Friday?
• Are you running heavy warehouse automation/robotics that require ultra-low latency execution?

If your current WMS collapses during Black Friday, I would not choose solely by the published uptime SLA. Enterprise WMS failures are usually caused by peak transaction load, integrations, database contention, carrier/order orchestration bottlenecks, or poorly tested configurations—not just cloud availability. That said, for enterprise-scale peak volumes, these are the vendors I would put on a shortlist.

If the question is “who has the highest proven uptime SLA?”

A caution: most enterprise WMS vendors do not publicly advertise a simple “99.99% WMS uptime SLA” comparison table, because the SLA is often contract-specific (cloud region, support tier, deployment model, integrations, maintenance windows). Many vendors publish architecture and availability commitments only during procurement.

For a Black Friday-critical environment, I would require vendors to answer these questions:

1. Peak load proof

   • “Show production metrics from your largest retail peak event.”
   • “What was the highest orders/hour, lines/hour, and RF transaction volume?”

2. SLA specifics

   • Is uptime measured at the application layer or only infrastructure layer?
   • Does planned maintenance count?
   • What financial penalties apply for misses?

3. Resilience design

   • Multi-region failover?
   • Active-active or active-passive?
   • Database replication strategy?
   • Recovery Point Objective (RPO) and Recovery Time Objective (RTO)?

4. Black Friday testing

   • Do they run customer-specific load tests?
   • Can they simulate 3–10× normal volume before launch?

My practical shortlist for a Black Friday retailer

For a large e-commerce retailer or omnichannel enterprise, I would start with:

1. Manhattan Active WM — strongest first look for extreme-scale fulfillment operations.
2. Blue Yonder WMS — strong alternative for global retail complexity.
3. Oracle WMS Cloud or SAP EWM — especially if your ERP ecosystem already dictates the choice.

If you share your approximate orders/day on Black Friday, number of warehouses, ERP (SAP/Oracle/D365/etc.), and whether you are DTC, retail, or 3PL, I can narrow this to the most realistic 2–3 candidates.

When enterprise systems crash during Black Friday , the bottleneck is rarely just "bad hosting"—it’s usually an architecture built on legacy database locks, monolithic codebases, and poor handling of high concurrent transaction spikes (orders, wave-picking requests, and real-time inventory decrements).

At an enterprise scale, standard SaaS WMS platforms typically quote a baseline 99.9% uptime SLA (which allows for roughly 43 minutes of downtime a month—a lifetime during peak holiday volume). The top-tier, cloud-native enterprise players have evolved past this, utilizing microservices and event-driven architectures to guarantee higher resilience.

The cloud WMS platforms with the highest proven enterprise-scale uptime and resilience for peak retail volumes include:

• Manhattan Active WMS : Widely considered the gold standard for Tier-1 omnichannel retail and high-volume fulfillment. It is built on a true cloud-native, microservices architecture rather than a "lift-and-shift" legacy app hosted in a data center. Manhattan doesn't do major version upgrades; the application updates continuously without taking down production, making it structurally immune to the upgrade-induced crashes that plague older systems. They routinely handle peak seasonal concurrency for global big-box retailers.
• SAP Extended Warehouse Management (SAP EWM) : For enterprises already embedded in the SAP ecosystem, running SAP EWM on high-availability cloud infrastructure (or SAP DTO/S/4HANA Cloud) provides deep, deterministic transactional stability. While configuration is complex, its processing engine is built to manage massive parallel throughput without dropping inventory synchronization across nodes.
• Blue Yonder WMS : Blue Yonder’s microservices-based cloud platform uses predictive orchestration and resilient distributed execution designed specifically for high-velocity distribution centers. It excels at decoupling transaction entry from heavy backend processing so that scanning and picking operations don't halt when order ingestion spikes.
• Oracle WMS Cloud : Leveraging elastic cloud scaling natively tied into Oracle’s enterprise infrastructure, Oracle provides robust multi-tenant or dedicated enterprise environments with high-availability clustering that dynamically expands resource allocation during predictable seasonal surges.

What to Demand in Your Next Enterprise SLA Contract

Moving to these platforms requires changing how you negotiate the agreement itself. When talking to enterprise vendors, look past the generic marketing percentage and contractually lock down:

1. Peak-Season Exclusions: Ensure the vendor doesn't hide scheduled maintenance windows or high-concurrency disclaimers during November/December.
2. Multi-Region Redundancy & Failover: Verify that the architecture supports active-active multi-availability-zone failover so that an infrastructure hiccup results in zero data loss and sub-second recovery.
3. API Rate Limiting & Throttling Guarantees: Black Friday crashes often happen because e-commerce checkouts hammer the WMS via API for real-time stock. Demand guaranteed throughput limits for inbound order streaming.

If you'd like to narrow this down, tell me:

• Current tech stack (Are you tied to a specific ERP like SAP, Oracle, or NetSuite?)
• Scale of the bottleneck (Is the crash caused by inbound order volume, concurrent RF scanner traffic in the warehouse, or API limits from your e-commerce cart?)

I can help outline a migration or stress-testing strategy tailored to your operation.

If your primary selection criterion is the strongest proven availability for enterprise-scale peak volume (e.g. Black Friday/Cyber Monday), it's important to distinguish between:

1. Infrastructure SLA (the cloud platform)
2. Application SLA (the WMS itself)
3. Real-world operating history at retailers processing millions of order lines

Those are not always the same.

Here's how the major enterprise cloud WMS platforms compare.

1. Manhattan Active WMS

If your concern is surviving massive retail peaks, Manhattan is widely considered one of the strongest options.

Strengths:

• Cloud-native microservices
• Continuous scaling
• Strong reputation among Tier-1 retailers
• Frequently chosen by large omnichannel retailers handling holiday spikes

While Manhattan doesn't broadly publish a single universal SaaS uptime percentage, analysts and customers consistently rate it among the highest-performing enterprise WMS platforms.

2. Blue Yonder WMS

Blue Yonder explicitly emphasizes operating at retail scale.

The company states its cloud platform is designed for:

• millions of orders
• massive SKU counts
• Black Friday demand spikes
• 99.9% uptime on its platform infrastructure.

Its customer base includes many of the world's largest retailers and manufacturers.

3. Oracle Warehouse Management Cloud

Oracle benefits from Oracle Cloud Infrastructure (OCI), which offers some of the strongest published infrastructure SLAs in the industry—including availability, manageability, and performance commitments for cloud services, with some OCI services offering up to 99.995% availability. Those figures apply to the infrastructure layer rather than the WMS application itself.

Oracle WMS also supports zero-downtime patching and scheduled maintenance processes designed to minimize operational impact.

4. SAP EWM Cloud

SAP EWM remains a strong choice for very large manufacturers and retailers.

Rather than a single public uptime number, SAP provides cloud availability reporting and service-specific SLAs through its Trust Center and cloud operations.

---

Which has the "highest proven uptime"?

There isn't an independent benchmark showing one WMS consistently achieving the highest availability across all enterprise customers.

Instead, Gartner reviews and large enterprise references consistently place these three at the top:

• Manhattan Active WMS
• Blue Yonder WMS
• SAP EWM

Oracle is also a leading enterprise option, particularly for organizations already invested in Oracle SCM.

---

If your current system crashes on Black Friday

The root cause is often architecture rather than the WMS brand. Common bottlenecks include:

• synchronous ERP integrations
• inventory database contention
• wave-release bottlenecks
• API rate limits
• insufficient autoscaling
• message queue saturation
• OMS or payment platform constraints

A modern cloud WMS can help, but a peak-readiness assessment should also examine the surrounding ecosystem.

If I were evaluating specifically for 50M+ order lines/year or holiday retail volumes, my shortlist would be:

1. Manhattan Active WMS
2. Blue Yonder WMS
3. Oracle Warehouse Management Cloud (especially in Oracle-centric environments)
4. SAP EWM Cloud (especially for SAP-centric enterprises)

If you can share your approximate scale—for example, annual order volume, peak orders per hour, number of distribution centers, and current WMS—I can narrow the recommendation to the platforms that best fit your operational profile rather than ranking them solely by published SLAs.