Analyze a job using logs

This document describes how to enable, generate, and view logs from Cloud Logging for a Batch job.

You can use logs to get information that is useful for analyzing your jobs. For example, logs can help you debug failed jobs.

Notably, logs are only generated after a job starts running and only if logging was enabled for the job. If you need to analyze a job without logs, view status events instead.

Before you begin

Enable logging for a job

To allow logs to be generated for a job, enable logs from Cloud Logging when you create the job:

  • If you create a job using the Google Cloud console, logs from Cloud Logging are always enabled.

  • If you create a job using the gcloud CLI or the Batch API, logs from Cloud Logging are disabled by default. To enable logs from Cloud Logging, include the following configuration for the logsPolicy field while creating the job:

    {
    ...
    "logsPolicy": {
        "destination": "CLOUD_LOGGING"
    }
    ...
}

Write and generate logs for a job

When logs from Cloud Logging are enabled for a job, Cloud Logging automatically generates any of the logs that are written for the job. Specifically, Batch jobs can have the following log types:

  • agent logs (batch_agent_logs): logs for activities from the Batch service agent.

    Batch automatically writes agent logs for every job that has enabled logging.

  • task logs (batch_task_logs): logs for any data that you've configured a job's runnables to write to the standard output (stdout) stream or standard error (stderr) stream.

    You can optionally write task logs for each job that has enabled logging.

View logs for a job

You can view a job's logs using the Google Cloud console, the gcloud CLI, Logging API, Go, Java, Python, or C++.

Console

To view a job's logs using the Google Cloud console, do the following:

  1. In the Google Cloud console, go to the Job list page.

    Go to Job list

  2. In the Job name column, click the name of a job. The Job details page opens.

  3. Click the Logs tab. Batch displays all logs associated with the job.

  4. Optional: To filter the logs do any of the following:

gcloud

To view logs using the gcloud CLI, use the gcloud logging read command:

gcloud logging read "QUERY"

where QUERY is a query for Batch logs that contains Batch filter parameters.

API

To view logs using the Logging API, use the entries.list method:

POST https://logging.googleapis.com/v2/entries:list
{
    "resourceNames": [
        "projects/PROJECT_ID"
    ],
    "filter": "QUERY"
    "orderBy": "timestamp desc"
}

Replace the following:

Go

Go

For more information, see the Batch Go API reference documentation.

To authenticate to Batch, set up Application Default Credentials. For more information, see Set up authentication for a local development environment. 

import (
	"context"
	"fmt"
	"io"

	batch "cloud.google.com/go/batch/apiv1"
	"cloud.google.com/go/logging"
	"cloud.google.com/go/logging/logadmin"
	"google.golang.org/api/iterator"
	batchpb "google.golang.org/genproto/googleapis/cloud/batch/v1"
)

// Retrieve the logs written by the given job to Cloud Logging
func printJobLogs(w io.Writer, projectID string, job *batchpb.Job) error {
	// projectID := "your_project_id"

	ctx := context.Background()
	batchClient, err := batch.NewClient(ctx)
	if err != nil {
		return fmt.Errorf("NewClient: %w", err)
	}
	defer batchClient.Close()

	adminClient, err := logadmin.NewClient(ctx, projectID)
	if err != nil {
		return fmt.Errorf("Failed to create logadmin client: %w", err)
	}
	defer adminClient.Close()

	const name = "batch_task_logs"

	iter := adminClient.Entries(ctx,
		// Only get entries from the "batch_task_logs" log for the job with the given UID
		logadmin.Filter(fmt.Sprintf(`logName = "projects/%s/logs/%s" AND labels.job_uid=%s`, projectID, name, job.Uid)),
	)

	var entries []*logging.Entry

	for {
		logEntry, err := iter.Next()
		if err == iterator.Done {
			break
		}
		if err != nil {
			return fmt.Errorf("unable to fetch log entry: %w", err)
		}
		entries = append(entries, logEntry)
		fmt.Fprintf(w, "%s\n", logEntry.Payload)
	}

	fmt.Fprintf(w, "Successfully fetched %d log entries\n", len(entries))

	return nil
}

Java

Java

For more information, see the Batch Java API reference documentation.

To authenticate to Batch, set up Application Default Credentials. For more information, see Set up authentication for a local development environment. 

import com.google.cloud.batch.v1.Job;
import com.google.cloud.logging.v2.LoggingClient;
import com.google.logging.v2.ListLogEntriesRequest;
import com.google.logging.v2.LogEntry;
import java.io.IOException;

public class ReadJobLogs {

  public static void main(String[] args) throws IOException {
    // TODO(developer): Replace these variables before running the sample.
    // Project ID or project number of the Cloud project hosting the job.
    String projectId = "YOUR_PROJECT_ID";

    // The job which logs you want to print.
    Job job = Job.newBuilder().build();

    readJobLogs(projectId, job);
  }

  // Prints the log messages created by given job.
  public static void readJobLogs(String projectId, Job job) throws IOException {
    // Initialize client that will be used to send requests. This client only needs to be created
    // once, and can be reused for multiple requests. After completing all of your requests, call
    // the `loggingClient.close()` method on the client to safely
    // clean up any remaining background resources.
    try (LoggingClient loggingClient = LoggingClient.create()) {

      ListLogEntriesRequest request = ListLogEntriesRequest.newBuilder()
          .addResourceNames(String.format("projects/%s", projectId))
          .setFilter(String.format("labels.job_uid=%s", job.getUid()))
          .build();

      for (LogEntry logEntry : loggingClient.listLogEntries(request).iterateAll()) {
        System.out.println(logEntry.getTextPayload());
      }
    }
  }
}

Python

Python

For more information, see the Batch Python API reference documentation.

To authenticate to Batch, set up Application Default Credentials. For more information, see Set up authentication for a local development environment. 

from __future__ import annotations

from typing import NoReturn

from google.cloud import batch_v1
from google.cloud import logging


def print_job_logs(project_id: str, job: batch_v1.Job) -> NoReturn:
    """
    Prints the log messages created by given job.

    Args:
        project_id: name of the project hosting the job.
        job: the job which logs you want to print.
    """
    # Initialize client that will be used to send requests across threads. This
    # client only needs to be created once, and can be reused for multiple requests.
    log_client = logging.Client(project=project_id)
    logger = log_client.logger("batch_task_logs")

    for log_entry in logger.list_entries(filter_=f"labels.job_uid={job.uid}"):
        print(log_entry.payload)

C++

C++

For more information, see the Batch C++ API reference documentation.

To authenticate to Batch, set up Application Default Credentials. For more information, see Set up authentication for a local development environment. 

#include "google/cloud/batch/v1/batch_client.h"
#include "google/cloud/logging/v2/logging_service_v2_client.h"
#include "google/cloud/location.h"
#include "google/cloud/project.h"

  [](std::string const& project_id, std::string const& location_id,
     std::string const& job_id) {
    auto const project = google::cloud::Project(project_id);
    auto const location = google::cloud::Location(project, location_id);
    auto const name = location.FullName() + "/jobs/" + job_id;
    auto batch = google::cloud::batch_v1::BatchServiceClient(
        google::cloud::batch_v1::MakeBatchServiceConnection());
    auto job = batch.GetJob(name);
    if (!job) throw std::move(job).status();

    auto logging = google::cloud::logging_v2::LoggingServiceV2Client(
        google::cloud::logging_v2::MakeLoggingServiceV2Connection());
    auto const log_name = project.FullName() + "/logs/batch_task_logs";
    google::logging::v2::ListLogEntriesRequest request;
    request.mutable_resource_names()->Add(project.FullName());
    request.set_filter("logName=\"" + log_name +
                       "\" labels.job_uid=" + job->uid());
    for (auto l : logging.ListLogEntries(request)) {
      if (!l) throw std::move(l).status();
      std::cout << l->text_payload() << "\n";
    }
  }

Filter Batch logs

You can filter for Batch logs by writing a query that includes one or more of the following filter parameters and zero or more boolean operators (AND, OR and NOT).

  • To filter for logs from a specific job, specify the job's unique ID (UID):

    labels.job_uid=JOB_UID

    where JOB_UID is the UID of the job. To get a job's UID, view the details the job.

  • To filter for a specific type of Batch logs, specify the log type:

    logName=projects/PROJECT_ID/logs/BATCH_LOG_TYPE

    Replace the following:

    • PROJECT_ID: the project ID of the project that you want to view logs for.
    • BATCH_LOG_TYPE: the type of Batch logs you want to view, either batch_task_logs for task logs or batch_agent_logs for agent logs.

  • To filter for logs with custom status events, specify that the log must define the jsonPayload.batch/custom/event field:

    jsonPayload.batch"/"custom"/"event!=NULL_VALUE

  • To filter for logs of one or more specific severities, specify the following comparison:

    severityCOMPARISON_OPERATORSEVERITY_ENUM

    Replace the following:

For more filter options, see the Cloud Logging query language documentation.

What's next