The problem you're facing
Without monitoring, you learn about incidents from your users: the site is down, the app is slow, billing failed. Every incident becomes a crisis, because you discover it at the worst time : when it already has a business impact.
Our answer
We set up complete monitoring: metrics (Prometheus, Grafana), centralised logs (Loki), graduated alerting and readable dashboards. Then we operate it: alerts come to us, with intervention procedures : not into a channel nobody watches.
What LaMeDuSe takes care of
- Metric monitoring: CPU, RAM, disk, network, services
- Application monitoring: response times, errors, availability
- Centralised logging (Loki, ELK)
- Graduated alerting with argued thresholds
- Grafana dashboards readable by your teams
- Documented intervention procedures
- Optional on-call: we receive alerts and intervene
Our methodology
- 1
Audit & mapping
Complete inventory of your information system: servers, network, backups, access rights, technical debt.
- 2
Takeover plan
Risk prioritisation, documentation of the existing setup and monitoring in place from the first week.
- 3
Operations & automation
Day-to-day administration, automation of repetitive tasks, security patches and backup management.
- 4
Continuous improvement
Monthly review of incidents, capacity and costs, with a shared improvement roadmap.
Relevant technologies
Security
Hardened configurations, privileged access management, centralised logging, managed firewalls and a tested 3-2-1 backup plan.
Hosting & infrastructure
We administer your existing infrastructure as well as our own cloud: the people and the infrastructure can come from the same group.
Maintenance
Contracts with committed response times, 24/7 on-call for critical environments and a monthly operations report.
Why LaMeDuSe
Our own European infrastructure is monitored with the same tools we deploy for clients: Prometheus, Grafana, Loki. We live daily with the alerts we install for you : we know what deserves to wake someone up, and what can wait until Monday.
Frequently asked questions
Monitoring collects and displays; supervision adds interpretation and action: argued thresholds, graduated alerts, intervention procedures. A metric without a threshold is a chart, not supervision.
Prometheus and Grafana for metrics, Loki or ELK for logs, active probes for availability. Proven open-source tools, with no per-server licence cost.
Yes: audit of the existing setup (what's watched, what's missing, what alerts for nothing), then progressive catch-up. Monitoring that alerts too much becomes noise nobody listens to : we fix that too.
Depending on the formula: your teams, our on-call teams, or both with a clear split by severity. What matters is that every alert has a responsible recipient and an associated procedure.