Configure log sinks

You can integrate your preferred log aggregators and observability platforms with your Northflank account.

Log sinks allow you to forward logs from containers in specific projects, or your whole account, allowing you to unify and process logs from your entire operation. You can improve your observability with the features of your preferred log service; analyze and visualise metrics, perform searches, set up real-time alerts, and meet your log auditing requirements.

You can also integrate with your own logging solutions by sending logs to a custom HTTP endpoint or an Amazon S3 bucket.

Northflank currently supports the following endpoints for log sinks:

• Datadog
• Loki
• Papertrail
• Mezmo (formerly LogDNA)
• Logtail
• Honeycomb
• AWS S3
• HTTP (custom endpoint)

By default, logs will be sent as plain text, some log sinks allow you to send logs as JSON instead.

Text

Text logs are sent as a single string containing only the log message itself in the format:

<timestamp: [yyyy-MM-dd’T’HH:mm:ss.SSSSSSSSS’Z’]> <stream [stdout | stderr]> <log line>.

For example:

2022-09-12T09:21:40.817924799Z stdout F Mon Sep 12 09:21:40 UTC 2022: Log item 138051

JSON

JSON logs are sent as an array of objects, with each object representing a single log entry:

  {
    [
      {
        "entity":"<entity_id>",
        "host":"Northflank",
        "message":"<datetime with timezone> <log message>",
        "path":"/",
        "pod":"<service_name>-<pod_id>",
        "project":"<project_name>",
        "service":"<service_name>",
        "source":"Northflank",
        "source_type":"http",
        "timestamp":"<datetime with timezone>",
        "type":"<service|job|addon>"
      },
      {...},
      {...}
    ]
  }

Log sinks can be created on your user or team account in your account settings.

To add a new log sink, select the type of log sink you want and follow the specific instructions below for each type.

You can pause and resume your log sinks as required after they have been created, update the authentication details, and change the projects they target.

Project restrictions

You can enable make log sink target selected projects to select specific projects to forward logs from. Only containers running in the selected projects will have their logs forwarded.

You can leave this disabled to forward logs from all running containers in all projects on your account.

If you are creating a log sink on a team account and your role is restricted to specific projects you will only be able to create log sinks targeting the projects you have access to.

Validation log

When you create or update a log sink a validation log is sent to the endpoint to ensure the credentials entered are correct. This validation log will be stored in your sink and takes the following format:

{
  "app-name": "log-forwarder",
  "host": "Northflank",
  "message": "Validating log sink credentials",
  "severity": "info",
  "date": "<datetime with timezone>"
}

1. Select the Datadog sink type in your new Northflank log sink
2. Select the region your Datadog account exists in by identifying it from the URL that you use to access it:

3. Navigate to organisation settings, open API keys, and create a new key. Make sure your account has at least the standard role, or a custom role with permissions to write logs .
4. Copy the API key and paste it into the field in Northflank
5. Select your project restrictions, if any, and add log sink
6. You should receive a validation log entry to confirm your credentials are valid, and logs will now be sent to the sink

You can send your logs to a self-hosted instance of Loki, or to a Loki instance hosted on Grafana cloud .

Grafana cloud

1. Select the Loki sink type in your new Northflank log sink
2. Navigate to your Grafana cloud stack
3. View the details for Loki
4. Follow the instructions to add Loki as a data source in your Grafana instance, if it's not currently configured. Copy and/or save your API key, as it will not be shown again.
5. Go to Northflank and enter the information from your Loki configuration:
   • URL
   • user
   • the API key you created when integrating Loki with Grafana
6. Choose the encoding method for the logs: text (default) or JSON
7. Select your project restrictions, if any, and add log sink
8. You should receive a validation log entry to confirm your credentials are valid, and logs will now be sent to the sink

Self-hosted

1. Select the Loki sink type in your new Northflank log sink
2. Enter the URL of your Loki deployment
3. Enter the username and password you set when configuring your Loki instance
4. Choose the encoding method for the logs: text (default) or JSON
5. Select your project restrictions, if any, and add log sink
6. You should receive a validation log entry to confirm your credentials are valid, and logs will now be sent to the sink

1. Select the Papertrail sink type in your new Northflank log sink
2. Navigate to your Papertrail settings and open log destinations
3. Create log destination, allow it to accept logs from unrecognised systems, and leave delete systems with no events enabled
4. Choose to accept connections via port (recommended) or token (HTTPS). Leave TLS encryption enabled for TCP, and plain text enabled for UDP for port connections.
5. On Northflank select the authentication method you chose when creating your log destination
   • For port enter the URL and port generated for your log destination, for example logs5.papertrailapp.com:12345
   • For token enter both the endpoint and token generated for your log destination
6. Choose the encoding method for the logs: text (default) or JSON
7. Select your project restrictions, if any, and add log sink
8. You should receive a validation log entry to confirm your credentials are valid, and logs will now be sent to the sink

1. Select the Mezmo (formerly LogDNA) sink type in your new Northflank log sink
2. In Mezmo, open settings, navigate to the API keys page under organisation
3. Generate a new ingestion key and copy the value
4. Return to Northflank and paste the ingestion key into the API key field
5. Select your project restrictions, if any, and add log sink
6. You should receive a validation log entry to confirm your credentials are valid, and logs will now be sent to the sink

1. Select the Logtail sink type in your new Northflank log sink
2. In your Logtail account navigate to sources and connect source
3. Name your source, select HTTP as the platform, and create source
4. Copy the source token, return to Northflank, and enter it into the API key field for your new log sink
5. Select your project restrictions, if any, and add log sink
6. You should receive a validation log entry to confirm your credentials are valid, and logs will now be sent to the sink

1. Select the Honeycomb sink type in your new Northflank log sink
2. In Honeycomb navigate to environments and manage environments
3. View API keys for the environment you want to use with Northflank and copy the API key value
4. Return to Northflank and paste the API key value into the form
5. Enter the name of your dataset, if it doesn't already exist in Honeycomb it will be created for you
6. Select your project restrictions, if any, and add log sink
7. You should receive a validation log entry to confirm your credentials are valid, and logs will now be sent to the sink

1. Select the Logz.io sink type in your new Northflank log sink
2. In Logz.io open your settings and select the manage tokens page
3. Open data shipping tokens and copy the default token, or create a new one
4. Return to Northflank and paste the API token into the form
5. If the listener URL displayed on the data shipping tokens page is different from the default listener.logz.io, copy it and replace the listener URL in the Northflank form
6. Select your project restrictions, if any, and add log sink
7. You should receive a validation log entry to confirm your credentials are valid, and logs will now be sent to the sink

You can create an S3 sink hosted by Amazon Web Services, or another implementation of S3 storage. The following instructions are to integrate S3 hosted by AWS, for other implementations the exact details may differ.

Logs sent to S3 buckets are automatically segmented into directories to help identify and find them, by project, resource type, and resource name. For example:

1. Select the AWS S3 sink type in your new Northflank log sink
2. Go to your AWS console and select S3 from the storage section in the services menu
3. Create a new bucket, or select an existing one, and enter the name of the bucket in Northflank
4. Enter the service endpoint, for example s3.eu-west-1.amazonaws.com for a bucket in the EU-West-1 region
5. Enter the region for your bucket, as indicated in the endpoint
6. Open the IAM console and create a new IAM user with the necessary permissions, or a role containing the necessary permissions
7. Copy the access key ID and access key secret to Northflank directly, and/or save them somewhere secure (you will not be able to view them again)
8. Choose whether to send uncompressed plaintext logs, or logs compressed with gzip to your bucket
9. Select your project restrictions, if any, and add log sink
10. You should receive a validation log entry to confirm your credentials are valid, and logs will now be sent to the sink

You can configure a custom HTTP endpoint to receive your logs from Northflank.

Log format

Your endpoint will receive logs from Northflank in either text or JSON format, and you can secure your endpoint by only accepting requests with either basic authentication or a bearer token.

Text logs will be sent with the header Content-Type: text/plain and the body will consist of a string containing the log line. JSON logs will be sent with the header Content-Type: application/json and the body will consist of a JSON object containing an array of log lines.

Authentication

Northflank support two types of HTTP authentication. For security your endpoint should require HTTPS, otherwise credentials and logs will be transmitted unencrypted.

• Basic authentication: basic auth requires a password, and optionally a username. This will be encoded using base64 before being sent in the header as Authorization: "Basic <encoded credentials>"
• Bearer token: your bearer token will be encoded using base64 and sent in the header of the request as Authorization: "Bearer <encoded token>"

Response

Your endpoint should return a 2xx response code.

Create your HTTP log sink on Northflank

1. Enter the URL of your log sink
2. Select your authentication strategy: basic authentication or bearer token
   • Enter your username and password to authenticate using basic authentication
   • Enter your token to authenticate using bearer token
3. Choose the encoding method for the logs: text or JSON
4. Select your project restrictions, if any, and add log sink
5. Your HTTPS endpoint should receive a validation log entry to confirm your credentials are valid, and logs will now be sent to the sink