Servers & Infrastructure

The SLA depends on the architecture: a single server without redundancy physically cannot guarantee 99.99% uptime. For demanding SLAs, we design for high availability (redundancy, load balancing, failover). What we always guarantee is the incident response time according to the contracted plan — from "next business day" to "15 minutes on 24/7 on-call."

We prefer access with elevatable privileges (sudo with MFA) instead of permanent root access. All access is logged in audit trails. In more sensitive environments, we work with a dedicated bastion host or Just-in-Time access with approval — whatever best fits your internal controls.

With 24/7 on-call plans, critical alerts trigger calls to on-call staff who respond within minutes. Standard plans handle critical issues during extended hours (6 AM - 10 PM) and non-critical issues during regular business hours. We clearly define what counts as "critical" from day one so there are no surprises.

AWS is the most mature and has the largest ecosystem, ideal for almost any case. Azure integrates naturally if you already use Microsoft 365, Active Directory, or Microsoft products. GCP stands out in analytics, big data, and ML. We evaluate this based on your current stack, team, and where the best cost/benefit ratio is for your workloads. Sometimes the answer is multi-cloud.

It depends on the architecture. Lift-and-shift typically costs similar to or more than on-premise. Real savings appear when you use cloud-native services (serverless, managed databases, auto-scaling) and optimize with reserved instances or savings plans. We perform a TCO (total cost of ownership) analysis with your real data before migrating so you go in with realistic expectations.

For stateless applications or those with replicable data, downtime can be a matter of minutes (DNS cutover). For large databases with heavy traffic, prior replication strategies and cutover allow for windows of 1-4 hours. For critical cases that can't have any downtime, dual-write or multi-region architectures are used, but the cost goes up. We define this based on criticality.

We keep two things separate: the cost of the VPS (billed directly by the cloud provider and held under your name) and our monthly management fee. That gives you transparency and control: if you ever decide to switch providers or take over management internally, you don't have to renegotiate anything.

For plans with 24/7 on-call enabled, monitoring triggers alerts to a rotating on-call engineer. For standard plans, critical incidents (site down, security breach) are handled immediately even on weekends; non-critical issues are queued for the next business day.

In most cases, yes. We prepare the new environment in parallel, sync the data, perform a low-TTL DNS cutover, and keep both active until everything is validated. The few cases with downtime are due to dependencies on the previous provider that don't allow for this; in those scenarios we coordinate a minimal maintenance window.

RPO (Recovery Point Objective): how much data you can acceptably lose, measured in time. If you back up every 4 hours, your RPO is 4 hours — that's the information you could lose in the worst case. RTO (Recovery Time Objective): how long you can be down before resuming operations. Both determine how costly and elaborate your DRP needs to be.

Both. Local backup (NAS, tape, another server) provides fast restoration for common failures. Cloud backup (S3, Glacier, Backblaze B2) protects against physical disasters (fire, theft, ransomware that also encrypts local backups). The 3-2-1 strategy combines both worlds — it's the recommended standard.

We restore to an isolated environment (sandbox, temporary VM) and validate that the data is intact and services start up correctly. No production data is touched. We document every test: what was restored, how long it took, and any issues that came up. That record is useful for audits and for calibrating real vs. theoretical RTOs.

Tier 1 handles routine incidents (password resets, basic usage issues, simple installations). Tier 2 diagnoses more complex technical problems (application failures, network configurations, performance issues). Tier 3 are specialists for critical incidents (security incidents, infrastructure failures, architecture). Each tier resolves what it can and escalates to the next when needed.

We offer all three models. Fixed monthly billing provides predictability (ideal for companies with steady volume). An hours bundle works well if volume is variable (you buy a block of hours consumed as needed). Per ticket is useful for occasional incidents. We recommend one based on your usage pattern in the initial proposal.

By default, extended business hours (Monday to Friday, 7:00 AM to 5:00 PM GMT-6). For plans with 24/7 on-call coverage, critical incidents are handled any day and time. We define with you what counts as critical (service down, security breach, inability to operate) so there's no ambiguity when an incident happens.

Both, but with clear rules. When we do consulting, our recommendation is objective — if the best solution is a product we don't sell, we say so. If it happens to match something we do implement, the decision to hire us for the execution is entirely yours, without pressure. Transparency about potential conflicts of interest is part of the process from the start.

Large consulting firms are ideal for multinational companies with complex processes and high budgets. For Guatemalan SMEs, our proposal is: direct senior team (not junior staff billed at senior rates), knowledge of the local market (vendors, compliance, real costs), and lower process overhead. The result tends to be more actionable and at a fraction of the cost.

We offer both. Some clients want only the assessment and roadmap to execute with their own team — perfect. Others want us to support the implementation of the priority initiatives. The decision is yours at every stage. The only thing we don't offer is "lock-in": if you decide to execute with another provider after the consulting engagement, the documentation is yours and clear enough to make that possible.

An outsourced technician responds when you call them — reactive, with no proactive visibility. An MSP operates with continuous monitoring tools, documented processes, objective metrics, and accountability for the entire environment. With MSP, we're responsible for uptime, not just for answering calls. It's a change of model, not just of provider.

Plans are designed with reasonable limits ("supported users," "managed servers," "included project hours"). Incidents within scope carry no additional cost — that's the essence of MSP. Larger projects (e.g., migrating an app to the cloud) are quoted separately because they fall outside the scope of daily operations. We define this clearly in the contract to avoid ambiguity.

Throughout the contract, we maintain complete documentation, written procedures, and version-controlled configurations. If at any point you decide to bring the function in-house, the handover is orderly: inventory, access credentials, runbooks, lessons learned. We support the transition during an agreed period. Our interest is for the relationship to work based on results, not lock-in.