Container Instances
An instance is a running copy of a container. It represents an actual process running on a single server.
Viewing Instance Output
Instances stream their console output in realtime back to the core. The console will store up to 128KB of console output between container restarts. For more history, refer to the documentation on container logs.
This stream can be viewed via the portal, via a websocket using the API, or via the CLI [coming soon].
Two-Way Console (SSH)
Cycle provides a service known as the Console Service, facilitates an SSH connection to the instance process, without installing SSH on the container instance or the server it runs on. This is critical for security, as these programs by their nature open up an attack surface.
Advanced Security
The Console Service acts as a secure encrypted proxy, located at https://console.cycle.io
. Credentials for accessing an instance via
the console service must be generated, and are only valid for 30 minutes. There are several advantages to this approach.
- No SSH is installed onto the server, or within a container instance. All communication is done via a secure encrypted tunnel back to Cycle's core.
- Generating credentials for access via SSH is restriced based on the role of the account. There is no need to generate or remember SSH keys. This is especially helpful when dealing with large teams.
- Credentials automatically expire after 30 minutes.
Generate Credentials
To generate credentials for accessing the two-way instance console, follow one of our interface-specific guides.
Migrating Instances
Instances can easily be migrated between servers, data centers, or even between cloud providers. Migrating an instance shuts down the instance and creates and starts a new instance on the target server.
Migrating Stateful Instances
When migrating a stateful instance with a volume attached, the instance is shut down and a tar.gz file is created off any attached volumes. The volume contents are migrated with the instance to the new server.
The original instance will not be deleted until the purge time is reached, in case the migration needs to be reverted. The purge time for stateful instances is 3 hours.
Reverting a Migration
Migrated instances can be reverted quickly. A reverted migration runs the same process as the original migration, in reverse.
Migrate Instances on Cycle
Telemetry
Live telemetry data (CPU, RAM, and network traffic) can be streamed directly from the instance process. Instance telemetry can be streamed in real-time, or as a report up to 6 hours by default, or up to 3 days when configured under the container deployment configuration.
Telemetry is stored on the instance's host machine for the retention period and will use some storage overhead.