Storing and Accessing Build Artifacts

This document describes how to work with Artifacts in the following sections:

Artifacts Overview

If your build produces persistent artifacts such as screenshots, coverage reports, core files, or deployment tarballs, CircleCI can automatically save and link them for you.

Find links to the artifacts at the top of the build page. Artifacts are stored on Amazon S3. There is a 3GB curl file size limit.

Artifacts are designed to be useful around the time of the build. It is best practice not to rely on artifacts as a software distribution mechanism with long term future guarantees.

Uploading Artifacts

To upload artifacts created during builds, use the following example:

version: 2
      - image: python:3.6.0

    working_directory: /tmp
      - run:
          name: Creating Dummy Artifacts
          command: |
            echo "my artifact file" > /tmp/artifact-1;
            mkdir /tmp/artifacts;
            echo "my artifact files in a dir" > /tmp/artifacts/artifact-2;

      - store_artifacts:
          path: /tmp/artifact-1
          destination: artifact-file

      - store_artifacts:
          path: /tmp/artifacts

The store_artifacts step uploads two build artifacts: a file (/tmp/artifact-1) and a directory (/tmp/artifacts). After the artifacts successfully upload, view them in the Artifacts tab of the build page in your browser. There is no limit on the number of store_artifacts steps a job can run.

Currently, store_artifacts has two keys: path and destination.

  • path is a path to the file or directory to be uploaded as artifacts.
  • destination (Optional) is a prefix added to the artifact paths in the artifacts API. The directory of the file specified in path is used as the default.

Uploading Core Files

This section describes how to get core dumps and push them as artifacts for inspection and debugging. The following example creates a short C program that runs abort(3) to crash the program.

  1. Create a Makefile with the following lines:

        gcc -o dump main.c
  2. Create a config.yml file with the following lines.

      #include <stdlib.h>
      int main(int argc, char **argv) {
  3. Run make and ./dump on the generated program to print Aborted (core dumped)!

Following is a full config.yml that compiles the example C abort program, and collects the core dumps as artifacts.

version: 2.0
      - image: gcc:latest
    working_directory: ~/work
      - checkout
      - run: make
      - run: |
    # tell the operating system to remove the file size limit on core dump files 
          ulimit -c unlimited
      - run:
          command: |
            mkdir -p /tmp/core_dumps
            cp core.* /tmp/core_dumps
          when: on_fail
      - store_artifacts:
          path: /tmp/core_dumps

The ulimit -c unlimited removes the file size limit on core dump files. With the limit removed, every program crash creates a core dump file in the current working directory. The core dump file is named core.%p.%E where %p is the process id and %E is the pathname of the executable. See the specification in /proc/sys/kernel/core_pattern for details.

Finally, the core dump files are stored to the artifacts service with store_artifacts in the /tmp/core_dumps directory.

Core Dump File in Artifacts Page

When CircleCI runs a build, a link to the core dump file appears under the Artifacts tab of the build summary.

Downloading All Artifacts for a Build on CircleCI

Use the following procedure to download your artifacts with curl.

  1. Create an API token in the application by going to User Settings > API Tokens.

  2. Click Create New Token, add a name in the dialog box and click Add API Token. Name tokens to help you remember where they are used.

  3. Copy the token string that appears in the table.

  4. CD to a directory where you would like the artifacts files to be downloaded and run the following command, copying in the token from Step 3:

export CIRCLE_TOKEN='?circle-token=:your_token'

curl$CIRCLE_TOKEN | grep -o 'https://[^"]*' > artifacts.txt

<artifacts.txt xargs -P4 -I % wget %$CIRCLE_TOKEN

Note 1: Replace all the variables above that start with a : with real values for your project (don’t include the colon).

Note 2: :vcs-type will be github or bitbucket.

Note 3: In the example, the xargs command runs four processes to download files in parallel. Adjust this value to your needs.

Explanation: The line beginning with curl fetches all the artifacts details for a build and pipes it through the grep command to extract just the URLs. The results are saved to the artifacts.txt file. Then, xargs reads the file and downloads each artifact to the current directory.