24hr Full-Stack Support

Full-stack support means one team covers everything your product depends on: the application code, the cloud it runs on and the AI services in between. Many hosting providers fix only the servers, and many development agencies only work office hours, so problems that fall between them wait until morning. Messy Works Night Watch covers the full stack with engineers on shift in the UK, Malta and Australia. There are three cover levels: business hours, out of hours (17:00 to 09:00 on weekdays, plus weekends and bank holidays), and 24/7, 365 days a year. Every level has a response time of 15 minutes or less.

With Night Watch, a Messy Works engineer does. The alert goes to the engineer on shift, who has your runbook (step-by-step fixes written for your product), read access to your code and infrastructure, and an agreed escalation route. The engineer restores service, keeps you updated while it's live, and writes up what happened. Your build team or agency gets a written handover the next morning. They're only woken if a fix needs product knowledge, and only at the thresholds you agreed in advance.

A support SLA (service level agreement) should include two things for each priority level. The first is a guaranteed response time: how long until a person is working on the issue. The second is a target resolution time: how long until it's fixed or has a stable workaround. Both should be reported every month. Messy Works Night Watch commits to:

Priority	What it means	Response	Target resolution
P1 Critical	Product down or unusable, or revenue or data at risk	15 minutes or less	2 hours
P2 Major	A core function broken, no workaround	1 hour	8 hours
P3 Minor	A fault with a workaround, or a question	4 hours	2 working days
P4 Request	A planned change, restore or access change	8 hours	By agreement

The 15-minute response applies at every cover level. All times are measured inside your chosen service hours.

See also the homepage FAQ: What does a 99.9% uptime SLA actually get me?

Response time is how long it takes for a person to start working on your issue. Resolution time is how long it takes to fix it or put a stable workaround in place. A fast response with no resolution target can still leave you down for hours, so check that a support contract includes both. Messy Works guarantees response times and sets target resolution times for every priority level. It reports how each incident performed against both every month.

Yes. This is exactly what Messy Works Night Watch was built for. Your agency keeps the build, the roadmap and in-hours support, and none of that is renegotiated. Messy Works takes the hours they aren't contracted for, typically 17:00 to 09:00 on weekdays plus weekends and bank holidays. There's a written handover at the start and end of each day, so nothing is discovered twice. Messy Works agrees with your agency what it can change during an incident. It restores service and records the cause, and anything touching the roadmap goes back to the agency to fix properly.

One person carrying a phone works until they're asleep, on holiday, ill or gone, and then customers find the outage first. A week has 168 hours, and a normal working week covers about 45 of them. Covering the rest properly takes enough people for a sustainable rota, monitoring that wakes the right person, and a written process. That's rarely worth building for a single product. Messy Works Night Watch gives you a staffed rota across three time zones for one monthly fee. Your developer stops being the last line of defence.

An AI product can fail while everything around it looks healthy. The page loads and the request returns, but the answer is wrong, slow or expensive. Messy Works Night Watch monitors the AI model as well as the infrastructure, covering:

• AI provider outages and rate limits, watched per endpoint with an agreed fallback.
• Slow generations and backed-up job queues, raised as incidents instead of being left to run.
• Token spend, alerted against a budget.
• Output quality and safety: failure rates, refusals and malformed responses, with a prompt and response trail kept for audit.

Without a fallback, an outage at your AI provider becomes an outage in your product. Messy Works agrees a fallback with you during onboarding. That could be a second model or provider, a cached or simpler response, or a clear message to users. It then watches each AI endpoint so the switch happens quickly. During the incident, the Messy Works engineer on shift keeps you updated and switches back once the provider recovers.

Messy Works Night Watch usually goes live two weeks after signing:

• Week 1: a discovery session with you and whoever built the product, agreed read access, and monitoring installed and tuned on your stack.
• Week 2: the runbook is written, the escalation route is agreed in writing, and an incident is rehearsed on day one of cover.

Your first monthly service report follows 30 days after cover starts. Onboarding runs alongside your normal work, and on your side it needs an introduction, an access approval and one sign-off.

Usually not. Cloud and hosting providers support their own platform: the servers, network and managed services they run. They don't fix your application code, your configuration or the AI services you've connected. Under the shared responsibility model (the cloud provider secures the cloud, and you secure what you put in it), those parts are yours. Messy Works Night Watch covers the application, the cloud setup and the AI services as one team. One incident needs one call, not two suppliers pointing at each other.

See also the homepage FAQ: Who is responsible when it breaks: us, the cloud provider, or you? and the Operate FAQ: Who owns our cloud accounts and code if Messy Works manages them?

Messy Works Night Watch is priced as one monthly fee per product. Two choices set the fee:

1. your cover level (business hours, out of hours or 24/7)
2. a monthly ticket allowance of 5, 10, 20, 30 or 50 tickets, where a ticket is one incident resolved or one service request, and unused tickets roll into the next month within the quarter

The fee is quoted after a discovery call, on a 12 or 24 month term agreed up front. Onboarding, monitoring setup and the runbook carry no separate charge. If you want to leave, there's no exit fee and the notice period is 60 days.

Messy Works is ISO 27001 certified, the international standard for information security management. Night Watch runs under certified controls for access, change management and incident management, and produces evidence your own auditors and insurers can rely on. Access to your systems is agreed during onboarding, limited to what's needed, and recorded. Every incident ends with a written review.