locomotive

A Railway sidecar service for sending webhook events when new logs are received. Supports Discord, Datadog, Axiom, BetterStack and more!

Configuration

Configuration is done through environment variables. See explanation and examples below.

Generic Webhook Log Formats

These are examples of what the body would look like in the POST request done by locomotive

For Plaintext logs

[
    {
        "_metadata": {
            "deploymentId": "577e2cf2-a1fc-4e0f-b352-2780bca73a94",
            "deploymentInstanceId": "bbdc4e76-7600-415f-9f6b-e425311cec51",
            "environmentId": "b5ce7ab5-96f1-4fa3-929b-fc883f89cbd1",
            "environmentName": "production",
            "projectId": "8a6502bf-6479-440c-a14f-78ecd52abf09",
            "projectName": "Railyard",
            "serviceId": "24335e07-e68b-498f-bc9e-1b9146436867",
            "serviceName": "Autorack"
        },
        "message": "Hello, World!",
        "level": "info",
        "severity": "info",
        "time": "2020-05-22T21:27:33Z",
        "_time": "2020-05-22T21:27:33Z",
        "dt": "2020-05-22T21:27:33Z",
        "datetime": "2020-05-22T21:27:33Z",
        "ts": "2020-05-22T21:27:33Z",
        "timestamp": "2020-05-22T21:27:33Z"
    }
]

For Structured JSON logs

[
    {
        "_metadata": {
            "deploymentId": "5b7c81b35-1578-4eb8-8498-44f4f517b263",
            "deploymentInstanceId": "46cda6d4-f76c-45cb-8642-c7265949e497",
            "environmentId": "b5ce7ab5-96f1-4fa3-929b-fc883f89cbd1",
            "environmentName": "production",
            "projectId": "8a6502bf-6479-440c-a14f-78ecd52abf09",
            "projectName": "Railyard",
            "serviceId": "55b1755f-f2c6-4f24-8d51-0ed3754b253e",
            "serviceName": "Superliner"
        },
        "level": "info",
        "severity": "info",
        "message": "Hello, World!",
        "example_string": "foo bar",
        "example_int": 12345678,
        "example_float": 1.2345678,
        "example_int_slice": [123, 456, 789],
        "example_string_slice": ["hello", "world"],
        "example_group": {
            "example_grouped_int": 12345678,
            "example_grouped_string": "Hello, World!"
        },
        "time": "2020-05-22T21:27:33Z",
        "_time": "2020-05-22T21:27:33Z",
        "dt": "2020-05-22T21:27:33Z",
        "datetime": "2020-05-22T21:27:33Z",
        "ts": "2020-05-22T21:27:33Z",
        "timestamp": "2020-05-22T21:27:33Z"
    }
]

Grafana Loki Plaintext Log Example

{
    "streams": [
        {
            "stream": {
                "deployment_id": "fb8172c8-a65d-48a4-9d1e-9d5ef986c9c3",
                "deployment_instance_id": "25dfeb9b-0097-4f91-820b-5dccc5009b1d",
                "project_id": "dce92382-c4e4-4923-bacd-3a5f7bcab337",
                "project_name": "Union Pacific Freight",
                "environment_id": "57d88ccb-8db9-4aef-957e-ecd94c41fdf8",
                "environment_name": "production",
                "service_id": "aa8ce660-dad0-4f7d-8921-46295d180c09",
                "service_name": "Dash 8",
                "severity": "error",
                "level": "error"
            },
            "values": [["1590182853000000000", "a plaintext message", {}]]
        }
    ]
}

Grafana Loki Structured Log Example

{
    "streams": [
        {
            "stream": {
                "deployment_id": "fb8172c8-a65d-48a4-9d1e-9d5ef986c9c3",
                "deployment_instance_id": "25dfeb9b-0097-4f91-820b-5dccc5009b1d",
                "project_id": "dce92382-c4e4-4923-bacd-3a5f7bcab337",
                "project_name": "Union Pacific Freight",
                "environment_id": "57d88ccb-8db9-4aef-957e-ecd94c41fdf8",
                "environment_name": "production",
                "service_id": "aa8ce660-dad0-4f7d-8921-46295d180c09",
                "service_name": "Dash 8",
                "severity": "error",
                "level": "error"
            },
            "values": [
                [
                    "1590182853000000000",
                    "hello, world",
                    {
                        "float": "10.51",
                        "number": "10",
                        "string_value": "hello world",
                        "user": "null"
                    }
                ]
            ]
        }
    ]
}

Notes:

Metadata is gathered once when the locomotive starts, If a project/service/environment name has changed, the name in the metadata will not be correct until the locomotive is restarted.
The body will always be a JSON array containing one or more log objects.
Various common timestamp attributes are included in every log object to increase compatibility with external logging services. ref 1, ref 2
The default Content-Type for these POST requests is set to application/json
Structured log attributes sent to Grafana Loki must always be a string

All variables:

RAILWAY_API_KEY - Your Railway API key.
- Project level keys do not work.
ENVIRONMENT_ID - The environment ID your service is in.
- Auto-filled to the current environment ID.
TRAIN - The ID of the service you want to monitor.
- Supports multiple service Ids, separated with a comma.
DISCORD_WEBHOOK_URL - The Discord webhook URL to send logs to.
- Optional.
DISCORD_PRETTY_JSON - Pretty print the RAW JSON object in Discord embeds.
SLACK_WEBHOOK_URL - The Slack webhook URL to send logs to.
- Optional.
SLACK_PRETTY_JSON - Pretty print the RAW JSON object in Slack embeds.
SLACK_TAGS - Tags to add to the Slack message.
- Supports multiple tags, separated with a comma.
- Optional.
LOKI_INGEST_URL - The Loki ingest URL to send logs to.
- Example with no authentication: https://loki-instance.up.railway.app/loki/api/v1/push
- Example with username/password authentication: https://user:[email protected]/loki/api/v1/push
- Optional.
INGEST_URL - The URL to send a generic request to.
- Example for Datadog: INGEST_URL=https://http-intake.logs.datadoghq.com/api/v2/logs
- Example for Axiom: INGEST_URL=https://api.axiom.co/v1/datasets/DATASET_NAME/ingest
- Example for BetterStack: INGEST_URL=https://in.logs.betterstack.com
- Optional.
ADDITIONAL_HEADERS - Any additional headers to be sent with the generic request.
- Useful for auth. In the format of a cookie. meaning each key value pair is split by a semi-colon and each key value is split by an equals sign.
- Example for Datadog: ADDITIONAL_HEADERS=DD-API-KEY=;DD-APPLICATION-KEY=
- Example for Axiom/BetterStack: ADDITIONAL_HEADERS=Authorization=Bearer API_TOKEN
REPORT_STATUS_EVERY - Reports the status of the locomotive every 5 seconds.
- Default: 5s.
- Format must be in the Golang time.DurationParse format
  - E.g. 10h, 5h, 10m, 5m 5s
LOGS_FILTER - Global log filter.
- Either ALL, INFO, ERROR, WARN or any custom combination of severity / level.
- Accepts multiple values, separated with a comma.
- Defaults to allowing all log levels.
- Optional.
LOGS_FILTER_DISCORD - Discord specific log filter.
- Same options and behavior as the global log filter.
LOGS_FILTER_SLACK - Slack specific log filter.
- Same options and behavior as the global log filter.
LOGS_FILTER_LOKI - Slack specific log filter.
- Same options and behavior as the global log filter.
LOGS_FILTER_WEBHOOK - Ingest URL specific log filter.
- Same options and behavior as the global log filter.

Log Filtering

You can filter logs by severity level and content using the following environment variables:

Level Filters

LOGS_FILTER: Global level filter applied to all outputs
LOGS_FILTER_DISCORD: Level filter applied to Discord output
LOGS_FILTER_SLACK: Level filter applied to Slack output
LOGS_FILTER_LOKI: Level filter applied to Loki output
LOGS_FILTER_WEBHOOK: Level filter applied to webhook output

Level filter options: ALL, INFO, ERROR, WARN, or any custom combination of severity / level.

Content Filters

LOGS_CONTENT_FILTER: Global content filter applied to all outputs
LOGS_CONTENT_FILTER_DISCORD: Content filter applied to Discord output
LOGS_CONTENT_FILTER_SLACK: Content filter applied to Slack output
LOGS_CONTENT_FILTER_LOKI: Content filter applied to Loki output
LOGS_CONTENT_FILTER_WEBHOOK: Content filter applied to webhook output

Content filters support regular expressions or plain text searches.

Examples:

"hello"
"[A-za-z]ello"