Public clusters
The shared servers that run our customers’ services–availability over 30 days
- Downtime
- –
- In total
- –
- Last 24 h
- –
- Active public clusters
- –
SLO –
SLA –
Error budget consumed–
Cloud in Algeria · DevOps · Software · AI · Data
The Algerian cloud platform that builds your software and runs it.
SkyVault is an engineering team based in Algeria. We build your applications and run them on ISOGrid, the Algerian cloud with the most managed services, and one of the simplest ways to put an application online: connect GitHub or GitLab, upload a .zip, or let the DevOps agent take care of it.
3 ways to deployThe console, the CLI or the DevOps agent: from GitHub, GitLab or a .zip
6 areas of expertiseCloud, software, DevOps, AI, data science, R&D
Cloud in AlgeriaLocal hosting, multiple providers, high availability
From 5,000 DZDMonthly subscription to the cloud platform
Live
The ISOGrid platform, in real numbers
These numbers are not marketing promises: they are measured continuously on our infrastructure and refreshed automatically every 30 seconds. No customer data is identifiable in them.
Monthly history, SLO and incidents →Web access
The addresses (subdomains and domains) served by our web servers–availability over 30 days
- Downtime
- –
- In total
- –
- Last 24 h
- –
- Measured since
- –
SLO –
Error budget consumed–
RPO of replicated data services0 s to 10 ms
The RPO is the maximum amount of data an outage can lose. On our data services with a replication policy, every write is copied continuously to other servers: if a server goes down, a replica takes over with, at worst, the last 10 milliseconds of writes missing, and most often with no loss at all. Services without replication are protected by automatic backups.
Registered users–Accounts created on the ISOGrid console
Services running–
Services deployed in total–Including services deleted since
Storage kept for our customers–
Understanding these numbers
- Availability
- The share of time during which a service was actually responding. We measure it every minute: a cluster is “available” if at least one of its control servers (managers) responds; a web address is available if its name resolves and our web servers answer on it. The percentage shown is the average of all measured minutes.
- Downtime
- The complement of availability: 100% minus availability. The “in total” figure converts it into minutes, as if the whole platform were a single service. For example, 0.1% downtime over 30 days is about 43 minutes.
- SLO (service level objective)
- The internal target our team sets for itself, stricter than what we promise contractually. It works as an alarm: as long as it is met, we ship new features; if it is at risk, reliability takes priority.
- SLA (service level agreement)
- The minimum availability we commit to for our customers. It covers the infrastructure (the clusters), because that is what we operate: an application that stops because of its own code is not our responsibility.
- Error budget
- The margin of downtime the SLO allows. With a 99.9% SLO, we can be unavailable 0.1% of the time, about 43 minutes over 30 days. The gauge shows how much of that margin has already been consumed; the objective is met as long as it does not exceed 100%.
- What counts as an unreachable address
- Every minute, we call each of the subdomains and domains our web servers serve. The address counts as reachable as soon as its name resolves and our servers answer, whatever the response: an error returned by the application itself is a matter of its configuration, not of our infrastructure. A custom domain whose DNS the customer has not yet configured is not counted.
- RPO (maximum data loss)
- The amount of data an outage can lose, expressed in time. On our data services with a replication policy, it ranges from 0 seconds to 10 milliseconds: every write is copied continuously to other servers, which take over the service if one of them goes down.
- Services running
- Everything our customers are running right now: applications, databases (dedicated instances and shared databases), managed services and solutions (Keycloak, Kafka, RabbitMQ, Redis, WordPress…), object storage and Kubernetes namespaces.
- Services deployed in total
- All services put online since the platform opened, including those their owners have deleted since. An application counts from its first deployment.
- Storage kept
- The storage space reserved and backed up for our customers’ services.
The simplest, and the most complete
As simple as a PaaS. As deep as a cloud.
Most platforms make you choose: the simplicity of a tool that deploys a Git repository, or the depth of a cloud that runs a real architecture. ISOGrid gives you both, in the same console, with the same CLI and the same agent.
Put an application online in three steps
01
Give us the code
Connect GitHub or GitLab in one click, from the console or from the terminal with isogrid connect github. No repository yet? Upload a .zip of the project folder.
02
The platform does the rest
It detects the port, builds the image and, if a Dockerfile is missing, the AI agent writes it and fixes the build until it passes.
03
Your application is online
Subdomain, TLS certificate, logs, metrics and automatic rollback on failure. A public HTTPS URL, right away.
Or tell the DevOps agent. It reads your repositories or your .zip, draws the architecture in front of you, tells you what is missing, proposes a priced plan and deploys when you approve it. If it lacks the code, it offers to connect GitHub or GitLab, or to upload a .zip, without leaving the conversation.
And when your system grows, nothing to relearn
Clustered PostgreSQL with 3 or 5 nodes, real-time replicas, distributed reads, automatic failover. Managed MySQL and MongoDB.
Kafka, RabbitMQ, Redis and Keycloak operated for you, connected to your applications through private networks.
Kubernetes and Docker Swarm, hourly namespaces or dedicated clusters, with us, in your public cloud or on your own servers.
Managed S3 object storage, hourly to monthly backups, automatic scaling.
High availability across several Algerian providers, with automatic failover in case of an incident.
A system of several services (frontend, API, workers, message queues, databases) deployed as a whole by the DevOps agent, with its architecture diagram.
Our services
One team, from code to production
Each service stands on its own. Combined, they cover the whole lifecycle of a digital product: designing it, building it, hosting it, operating it and evolving it.
ISOGrid
Cloud platform
Push your code, get an application online: build, deployment, managed databases, TLS and monitoring included. On our cloud or installed on your premises.
See the platform → Custom-builtSoftware development
Web and mobile applications, APIs, back offices, payment and authentication integrations. Delivered with tests, documentation and a path to production.
See the service → Architecture & hostingCustom cloud services
Architecture, migration, Algerian multi-cloud and high availability. Private or on-premises cloud for organizations that must keep their data in-house.
See the service → OperationsDevOps services
Continuous integration and deployment, containers, orchestration, observability, infrastructure as code and on-call. Your production becomes predictable.
See the service → Artificial intelligenceAI solutions
Agents, assistants and automations connected to your data and your tools, with hosted or private language models. Measured, secured, in production.
See the service → R&DAdvanced research & data science
Modeling, data pipelines, dashboards and applied research prototypes. When the question has no off-the-shelf answer yet.
See the service → Data protectionDPO and Law 18-07 compliance
Outsourced data protection officer, register of processing activities, impact assessments, declarations to the ANPDP and data breach procedure: we bring your company into compliance with Law 18-07 as amended by Law 25-11, on paper and in your systems.
See the service →
ISOGrid
The operations layer a platform team would cost you
You are not buying compute. You get everything that goes around it: building your images, deployment, routing, certificates, backups and monitoring. Available by subscription on isogrid.skyvault.pro, or installed on your own servers with ISOGrid Nomad. No lock-in: your private clusters remain standard Docker Swarm or Kubernetes, and keep running if you leave.
Deployment from GitHub, GitLab, a .zip or a Docker image registry, through the console, the CLI or the DevOps agent
AI agent that writes, builds and fixes your Dockerfiles
Managed clustered PostgreSQL with automatic failover
Subdomain, TLS certificate and routing configured on every deployment
Live logs, metrics and alerts
Organizations, roles and isolated private networks
Clusters in your public cloud (Google Cloud, Azure, AWS…), on VPS or on your own servers
Our method
How we work
A short scoping phase, frequent deliveries, and a production we remain responsible for after go-live.
01
Scoping
A conversation to understand the need, the context and the constraints. You receive a proposal priced per feature, with no surprises.
02
Build
Iterative development, code review, automated tests and documentation. You see the product move forward week after week.
03
Go-live
Deployment on ISOGrid or on your infrastructure, with monitoring, backups and high availability from day one.
04
Operations
Platform subscription and hourly support to evolve the product, fix things fast and stay online.
Why SkyVault
Built for Algerian startups and SMEs
Local hosting. Your data stays in Algeria. ISOGrid integrates with the infrastructures of Algerian operators and hosting providers, such as Djezzy and ICOSNET, and fails over automatically from one to another in case of an incident. Cloud in Algeria with ISOGrid.
The simplest and the most complete. An application online in three steps from GitHub, GitLab or a .zip, and the same platform runs clustered PostgreSQL, Kafka and Kubernetes when you need them. See how.
Local payment and messaging. Proven Chargily, SMS and WhatsApp integrations, with several providers for high availability.
A single point of contact. Development, hosting and operations by the same team, with no passing the buck.
Prices in dinars, no surprises. Fixed prices per feature, monthly platform subscription and support billed by the hour.
Contact
Let’s talk about your project
Write to us or call us. We answer in French, Arabic and English.
Sales, contracts & communication
Aziz Taleb
Ready to deploy?
Create your organization on ISOGrid and put your first application online today.