Webcurl Hosting Platform At Webcurl, we have been providing hosting for our customers’ websites and applications since inception, and take pride in maintaining our understanding of all levels of the stack that powers your site. Our offering is based on the Kubernetes orchestration platform, which has become the de-facto approach for managing containerised workloads across multiple machines. Works anywhere Kubernetes transforms any group of servers into a self-managing and fault-tolerant cluster. It can run on a wide variety of platforms, and is well supported by all major cloud infrastructure vendors (AWS, GCE, Azure, and many more). This allows us to deploy sites in a standard way whether you need an on-premise or cloud solution, and can even deploy to multiple providers to allow seamless recovery in the event of an outage. Scalable for any situation Whether it’s the busy time of year, or just time for your fifteen minutes of fame, the cluster architecture offers agility to rapidly scale and meet the challenge. Kubernetes can monitor key metrics (e.g. CPU utilization or request rate) and automatically act to ensure availability during busy periods. Resources can be re-balanced quickly – perhaps spinning up an additional copy of the site on a less-busy server. If supported by the underlying environment, the system can even provision additional virtual machines as temporary members of the cluster to handle demand. Resilient against failure By operating across multiple machines, services can be architected to avoid single points of failure. Website data is stored on a clustered database, which typically replicates across 3 server nodes, to provide high availability and improved performance. The system constantly watches the health status of all key components. If any given server becomes unresponsive, traffic will immediately be routed elsewhere. Failed services will automatically try to restart or migrate to other cluster nodes, which can often mitigate issues in the time before the server operator has finished logging in. Website uptime and health data is monitored, and anomalies are forwarded to our incident management platform, and our engineering team notified. Built-in security features By running applications within containers, we limit available system privileges and reduce the scope of damage a compromised application could cause. Network policies ensure that applications may only communicate with pre-approved services. Activity logs are stored in a centralised system, which can analyse and detect malicious actions. Simple management For organisations with many web properties, we can provide access to a control panel for deploying multiple sites and environments on the platform. Backup and Disaster Recovery Disaster can strike at any time, and it can take many forms. We work proactively to ensure that we have the right systems and processes ready to deal with whatever might happen, and ideally prevent major issues from occurring in the first place! A key component in our preparedness is maintaining suitable backups of all site code and data. These are encrypted and stored with an appropriate cloud provider using durable storage, with a rotation strategy to help ensure critical backups are not overwritten prematurely. Recovery Strategies Cluster configuration state is backed up regularly. This can be used to roll back to a known-good state, or to provision an entirely new cluster Application file and database data is backed up to another location externally Application code is stored within our GitLab system with a full change history. In the unlikely event of a catastrophic failure, these technologies and processes allow us to initiate a complete reinstallation, with the last known good database within hours. Testing the Plan The plan is tested at least once a year, and any issues discovered are addressed, and the plan adapted as necessary. We continuously monitor our backup systems for errors, and regularly test-restore data to ensure it is consistent.