|
| 1 | +# Configuration # |
| 2 | + |
| 3 | +This page provides a (partial) list of configuration options that can |
| 4 | +be specified in Serverledge configuration files. |
| 5 | +All the supported configuration keys are defined in `internal/config/keys.go`. |
| 6 | + |
| 7 | +## Configuration files |
| 8 | + |
| 9 | +You can provide a configuration file using YAML or TOML syntax. Depending on the |
| 10 | +chosen format, the default file name will be `serverledge-conf.yaml` or |
| 11 | +`serverledge-conf.toml`. The file can be either placed in `/etc/serverledge`, |
| 12 | +in the user `$HOME` directory, or in the working directory where the server is |
| 13 | +started. |
| 14 | + |
| 15 | +Alternatively, you can indicate a specific configuration file when starting the |
| 16 | +server: |
| 17 | + |
| 18 | + $ bin/serverledge <config file> |
| 19 | + |
| 20 | + |
| 21 | +## Frequently used options |
| 22 | + |
| 23 | +| Configuration key | Description | Example value(s) | |
| 24 | +| ------------- | ------------- | -----------------| |
| 25 | +| `etcd.address` | Hostname and port of the Etcd server acting as the Global Registry. | `127.0.0.1:2379` | |
| 26 | +| `api.port` |Port number for the API server. | 1323| |
| 27 | +| `cloud.server.url` |URL prefix for the remote Cloud node API. | `http://127.0.0.1:1326` | |
| 28 | +| `factory.images.refresh` |Forces function runtime container images to be pulled from the Internet the first time they are used (to update them), even if they are available on the host.| `true` | |
| 29 | +| `container.pool.memory` |Maximum amount of memory (in MB) that the container pool can use (must be not greater than the total memory available in the host).|4096| |
| 30 | +| `janitor.interval` |Activation interval (in seconds) for the janitor thread that checks for expired containers.| 60| |
| 31 | +| `container.expiration` |Expiration time (in seconds) for idle containers. | 600| |
| 32 | +| `registry.area` |Geographic area where this node is located.| `ROME`| |
| 33 | +| `registry.udp.port` |UPD port used for peer-to-peer Edge monitoring.|| |
| 34 | +| `scheduler.policy` |Scheduling policy to use. Possible values: `default`, `localonly`, `edgeonly`, `cloudonly`.|| |
| 35 | + |
| 36 | +<!-- TODO: |
| 37 | +| `container.pool.cpus` ||| |
| 38 | +| `cache.size` ||| |
| 39 | +| `cache.cleanup` ||| |
| 40 | +| `cache.expiration` ||| |
| 41 | +| `scheduler.queue.capacity` ||| |
| 42 | +| `metrics.enabled` ||| |
| 43 | +| `metrics.prometheus.host` ||| |
| 44 | +| `metrics.prometheus.port` ||| |
| 45 | +| `registry.nearby.interval` ||| |
| 46 | +| `registry.monitoring.interval` ||| |
| 47 | +| `registry.ttl` ||| |
| 48 | +--> |
0 commit comments