Last Updated: September 3, 2025

​
How to Log From Your Container

When running a container on SaladCloud, you can log to the standard output and standard error streams. These logs are collected and stored by SaladCloud, and exported to any configured external logging source. Learn more about logging best practices.

​
Basic Logging

​
How to View Container Logs In the Portal

Once your container is up and running on a SaladCloud node, you can view the logs from your container on the “Container Logs” tab of a Container Group. To access this feature in the portal, navigate to your Container Group details page, and click on the ‘Container Logs’ tab. You can also navigate to this page from the Action menu of an instance. This option will take you to the Container Logs tab with the Machine ID value pre-filled. The Container Logs feature has a simple interface meant to provide you with enough visibility for basic troubleshooting of your applications running on SaladCloud, but it is not an exhaustive logging solution and lacks analytics and filtering features that are available on dedicated logging platforms. For more robust logging functionality, we recommend configuring the External Logging Service option in your container group. The Container Logs feature allows you to query your container logs with these options:
  • search for specific log messages with case-insensitive string filtering
  • filter for logs from a specific Machine ID. If left blank, logs from all Machine IDs in the Container Group will be returned.
  • filter for logs in a specified time period. Choose one of the predefined options, or use the custom start date and end date option
When your query returns results you will see them in a scrollable table view. Each result includes:
  • Log Message
  • Machine ID (The unique ID of the node running the container that created the log message.)
  • Log Create Time
Should your query yield no results, the table view will display “No data to display” Availability and Limitations This feature is available exclusively through the SaladCloud portal interface, not accessible via the public API. It’s in an early “Preview” stage, with the following considerations:
  • A maximum of 500 logs can be returned per query. Utilize specific query filters and time ranges to get the logs you are looking for.
  • Logs are retained and accessible for up to 90 days.
  • If you encounter this error message, try narrowing your query’s time range

​
Advanced Logging

For more advanced on-platform logging capabilities, you can make use of the Query Log Entries endpoint in the API. This endpoint, rather than being scoped at the container group level, is scoped at the organization level, allowing you to view and aggregate logs from any number of container groups and projects. This endpoint supports a more flexible query language as well, allowing you to make much more sophisticated filters across many fields, and supporting filters on structured logs emitted by your application. See the full query reference. You can query logs by making a POST request to the query logs endpoint like this: 
curl --request POST \
  --url https://api.salad.com/api/public/organizations/$organization_name/log-entries \
  --header 'Content-Type: application/json' \
  --header "Salad-Api-Key: $salad_api_key" \
  --data '{
  "end_time": "2023-11-07T05:31:56Z",
  "page_size": 1,
  "query": "<string>",
  "sort_order": "desc",
  "start_time": "2023-11-07T05:31:56Z"
}'
and receive a response like: 
{
  "items": [
    {
      "json_log": {},
      "parent_span_id": "<string>",
      "receive_time": "2023-11-07T05:31:56Z",
      "resource": {
        "labels": {},
        "type": "<string>"
      },
      "severity": "debug",
      "span_Id": "<string>",
      "time": "2023-11-07T05:31:56Z",
      "trace_Id": "<string>"
    }
  ],
  "organization_name": "acme-corp",
  "page_max_time": "2023-11-07T05:31:56Z",
  "page_min_time": "2023-11-07T05:31:56Z"
}
Each log event takes the following structure. 
{
  "time": "string iso8601",
  "receive_time": "string iso8601",
  "resource": {
    "type": "string enum",
    "labels": {
      // key-value pairs (string to string) specific to `type`
    }
  },
  "severity": "string enum",
  "severity_number": "integer",

  "trace_id": "string", // optional
  "span_id": "string", // optional
  "parent_span_id": "string", //optional

  // only one of `text_log` or `json_log`
  "text_log": "string",
  "json_log": {
    // key-value pairs (string to any valid JSON type) specific to log entry
  }
}
  • Note that this will contain EITHER text_log OR json_log, but not both. If you emit logs that are valid json, they will end up in the json_log field. Otherwise, the full log line will be in the text_log field.
  • The time field indicates the system time on the node where the log was emitted.
  • the receive_time field indicates the time when Axiom (what we use under the hood for log storage) received the log entry.
  • Any of these fields can be used in the query language. Here are some examples of queries you can make. These queries can be passed as the query parameter in the body of the request to the API.