We are operating a public dashboard, that shows an outside view of various parts of SWITCHengines and their availability or performance. The dashboard also allows you to subscribe to mails that will alert you to any anomalies with the system.
We are measuring various different use cases.
There are two indicators here:
- "Identity - Keystone" measures the availability of the central Identity service of SWITCHengines. It is used for every request that accesses the control plane. If the identity service is not available, operations like starting / stopping VMs, adding users or projects and the Web user interface become unavailable. We strive for > 99.95% availability
- "Web UI" simulates a user working with the Web Interface of SWITCHengines. Logging in, listing virtual machines, images, volumes and networks. This test touches all the API endpoints of all services that make up SWITCHengines. If it is not available or slow, users cannot manage their virtual machines. We strive for >99.95% availability and total run time of the test <30seconds
SWITCHengines API ZH/LS
SWITCHengines consists of various services that have different responsibilities. In this group, we measure the availability of the various API endpoints.
- Compute - Management of Virtual Machines
- Images - Management of Disk Images and Snapshots
- Volumes - Management of Disk Volumes
- Network - Management of virtual network functions
- ObjectStore - Access to Object Store
We strive for >99.95% availability of all API endpoints.
All network traffic in and out of SWITCHengines passes so called "Network Nodes". They handle the work of connecting the outside world with the various virtual machines. Each region has several network nodes, and the customer VMs (actually the private virtual routers of the various projects) are distributed among these network nodes. If a network node goes down, all VMs that are being serviced by that node will loose connectivity. The indicators in this group measure the connectivity from the outside to a virtual machine behind each of theses nodes. We strive for >99.95% availability here.
Our S3 compatible object store allows users to store and retrieve data. The indicators here measure that access is available and the speed of creation of objects in the object store.
We strive for >99.95% availability
SWITCHengines Supporting Applications
We have various applications that help us run SWITCHengines. Their availability doesn't impact the running of the service, but of functions like registration or billing.
- EnginesAdmin - the administrative user interface that allows customers to create and manage projects and users. We measure the availabiliyt of the Webserver of this application. We strive for >99.95% availability