REFERENCE / GLOSSARY
Glossary
Terms to use correctly in front of a customer. Each definition is taken from Fly.io’s docs or blog, with the source beside it.
This page is drawn from https://fly.io/docs/getting-started/essentials/, https://fly.io/docs/apps/overview/, https://fly.io/docs/machines/overview/, https://fly.io/docs/reference/fly-proxy/, https://fly.io/docs/networking/, https://fly.io/docs/volumes/overview/, https://fly.io/docs/sprites/ and the Fly.io blog.
Platform terms
| Term | Meaning | Source |
|---|---|---|
| Fly.io platform | “All the primitives, products, and features that make up the Fly.io public cloud.” | https://fly.io/docs/getting-started/essentials/ |
| Fly App | “A Fly App is an abstraction for a group of Fly Machines running your code on Fly.io.” It also holds config, Anycast IPs, certificates, secrets and optional volumes. | https://fly.io/docs/apps/overview/ |
| Organization | The administrative entity that manages billing, member access and shared environments. | https://fly.io/docs/getting-started/essentials/ |
| Machine | A Firecracker microVM. “Fly Machines are fast-launching virtual machines (VMs) that are individually runnable and controllable.” | https://fly.io/docs/getting-started/essentials/ |
| Fly Launch | Fly.io’s orchestrator for app hosting: fly launch, fly.toml, fly deploy, fly scale. | https://fly.io/docs/getting-started/essentials/ |
| flyctl | The Fly CLI, run as fly. It runs on your device. | https://fly.io/docs/getting-started/launch-demo/ |
fly.toml | The app configuration file: builds, env vars, services, mounts and release commands. | https://fly.io/docs/reference/configuration/ |
| Region | A location where Machines run, named by a code such as iad. fly platform regions lists them. | https://fly.io/docs/reference/regions/ |
Components you may hear about
| Term | Meaning | Source |
|---|---|---|
| flyd | The Fly Machines orchestrator that runs on the host. “Fly Apps run on flyd.” | https://fly.io/blog/design-and-implementation/, https://fly.io/docs/apps/overview/ |
| flaps | The Machines API server: “we call it flaps, but you don’t have to care”. It picks a worker server in the region. | https://fly.io/docs/machines/overview/ |
| Flaps (logs) | A different thing in a 2023 post: the proxy in front of NATS that “ensures you only see your own logs”. Check which one is meant. | https://fly.io/blog/shipping-logs/ |
| Fly Proxy | “the omnipresent routing layer between the public Internet and the Fly.io platform”. It does load balancing, handlers, backhaul and autostop/autostart. | https://fly.io/docs/reference/fly-proxy/ |
| corrosion | The service catalog Fly Proxy reads for apps, Machines and services, with gossip-based discovery. | https://fly.io/docs/reference/fly-proxy/ |
Networking terms
| Term | Meaning | Source |
|---|---|---|
| Anycast | “We announce global IP blocks from all of our datacenters over BGP, otherwise known as Anycast.” | https://fly.io/docs/networking/services/ |
| 6PN | “Fly Apps in an organization are connected by a mesh of WireGuard tunnels using IPv6 called a 6PN.” | https://fly.io/docs/networking/private-networking/ |
.internal | Fly.io DNS under which every app in the org has a name, such as my-app-name.internal. Only started Machines are returned. | https://fly.io/docs/networking/private-networking/ |
| Flycast | “A Flycast address is an app-wide private IPv6 address that Fly Proxy can route to over the private network.” | https://fly.io/docs/networking/flycast/ |
fly-replay | A response header that makes Fly Proxy replay the request to another region, Machine or app. | https://fly.io/docs/networking/dynamic-request-routing/ |
| Autostop/autostart | Fly Proxy stopping and starting existing Machines by traffic. It never creates or destroys Machines. | https://fly.io/docs/launch/autostop-autostart/ |
| Suspend | A Machine state that snapshots memory, so resume takes “a few hundred ms”. | https://fly.io/docs/reference/suspend-resume/ |
Storage and data terms
| Term | Meaning | Source |
|---|---|---|
| Volume | Local persistent storage for Machines: a slice of an NVMe drive on one server, one Machine at a time, one app. | https://fly.io/docs/volumes/overview/ |
| MPG | Managed Postgres: “our fully-managed database service”, with backups, high availability and failover. Only some regions host it. | https://fly.io/docs/postgres/, https://fly.io/docs/reference/regions/ |
| Tigris | “a globally distributed, multi-cloud object storage service with built-in support for the S3 API”. Billed at the provider’s list price. | https://fly.io/docs/tigris/, https://fly.io/docs/about/pricing/ |
Sprites terms
| Term | Meaning | Source |
|---|---|---|
| Sprite | “Sprites are persistent, hardware-isolated Linux environments for running arbitrary code.” | https://fly.io/docs/sprites/ |
| Warm and cold | Warm: VM suspended, memory kept, wake in 100 to 500ms. Cold: VM stopped, memory dropped, wake in 1 to 2s. | https://fly.io/docs/sprites/concepts/lifecycle/ |
| Checkpoint | A save point you create on purpose: a snapshot of the writable filesystem. It excludes running processes and memory. | https://fly.io/docs/sprites/concepts/checkpoints/ |
| Restore | Replaces the writable overlay with a checkpoint and restarts the environment. It is destructive. | https://fly.io/docs/sprites/concepts/checkpoints/ |
| Connector | Stores a credential once in your organization and routes API calls through the Sprites gateway, so Sprites never see the token. | https://fly.io/docs/sprites/concepts/connectors/ |
Word-choice reminders
- Say “Machine” for a VM and “app” for the group. A volume belongs to an app and attaches to one Machine (https://fly.io/docs/volumes/overview/).
- Do not describe a Sprite as just a Machine. The Sprites post says “Sprites are something new” and compares them to disposable ballpoint pens (https://fly.io/blog/design-and-implementation/).