Knapsack Pro

Bitbucket Pipelines vs Travis CI comparison of Continuous Integration servers
What are the differences between Bitbucket Pipelines and Travis CI?

Bitbucket Pipelines

https://bitbucket.org/product/features/pipelines

Travis CI

https://travis-ci.org
Unique feature

Best Jira integration possible

Build Matrix, ease of use, GitHub integration

Type of product

SaaS / On Premise

SaaS, Self-hosted / On Premise

Offers a free plan

Yes

Offers a very modest free cloud plan, limited to 5 users, 50 minutes of build time per month and 1GB storage. There's no free self-hosted version, but they do offer a $10 one-time payment plan for 10 users (build time and storage is only limited by your infrastructure)

Yes

Free for open source projects

Predictable pricing

Yes

Pricing is based on amount of users for both the cloud and on premise versions. The cloud offering has different tiers depending on build times and storage.

Yes

Clearly defined monthly plans, depending on concurrent jobs needed.

Support / SLA

Yes

Dedicated tehnical support.

Yes

Available via email, or dedicated online interface for paid plans.

Paralellism
Every CI servers tends to address this differently (parallel, distributed, build matrix). Some of it is just marketing, and some is just nuance. For this table, parallel means that tasks can be run concurrently on the same machine, distributed means that tasks can be scaled horizontally, on multiple machines
How to split tests in parallel in the optimal way with Knapsack Pro

Yes

Yes

TravisCI makes it very easy to split your build into different stages which are then run in parallel (ie: run integration tests separate from the unit tests). TravisCI calls this a build matrix: https://docs.travis-ci.com/user/build-matrix/. You can also very easily split tests accross several VMs using the knapsack_pro gem.

Distributed builds
distributed means that tasks can be scaled horizontally, on multiple machines
How to split tests in parallel in the optimal way with Knapsack Pro

N/A

Documentation is unclear, but it's reasonable to assume that distributed builds for the on premise version are not an issue.

N/A

Containers support / Build environment

Yes

Yes

TravisCI runs each build in a isolated virtual machine. Pre-build packages include a few which support specific languages (Ruby and JavaScript included) or other software (Git, various databases), but vanilla packages such as Ubuntu Trusty are also available.

Analytics / Status overview
Analytics and overview referrs to the ability to, at a glance, see what's breaking (be it a certain task, or the build for a specific project)

Yes

Excellent overview and contextual feedback.

Yes

Available by default in Travis (this is what most of the web UI consists of)

Management support
How easy is it to manage users / projects / assign roles and permissions and so on

Yes

N/A

Self-hosted option

Yes

Yes

Hosted plans / SaaS

Yes

Yes

Build pipelines
A continuous delivery pipeline is a description of the process that the software goes through from a new code commit, through testing and other statical analysis steps all the way to the end-users of the product.

Yes

No

Specifically built around GitHub pull requests. Pipelines can be defined, but parts of the process need to be implemented separatelly in GitHub.

Reports
Reports are about the abilty to see specific reports (like code coverage or custom ones), but not necesarily tied in into a larger dashboard.

Yes

Yes (partial)

No persistent storage eliminates the possibility of code coverage reports on TravisCI alone. There is support for integrated 3rd parties such as Coveralls for reporting code coverage.

Ecosystem
Besides the official documentation and software, is there a large community using this product? Are there any community-driven tools / plugins that you can use?

Yes

Large collection of available apps: https://confluence.atlassian.com/bitbucket/apps-and-integrations-675189068.html

No plugin support in TravisCI, plugins for other tools

Specific language support: Ruby
Some CI servers have built-in support for parsing RSpec or Istanbul output for example and we mention those. Some others make it even easier by detecting Gemfiles or package.json and automate parts of the process for the developer.

Yes

Clear, concise documentation on setting up a Ruby project with Bitbucket pipelines: https://confluence.atlassian.com/bitbucket/ruby-with-bitbucket-pipelines-872005618.html

Yes

TravisCI is designed to be a simple way to integrate CI/CD in your workflow so it has a couple of features aimed at specific languages, such as Ruby, starting from pre-built containers (with RVM already installed, for example) all the way to automatically running specific platform commands (such as detecting a Gemfile in the root of the project and automatically bundling dependencies). TravisCI also builds a Ruby SDK for easier use of the API.

Specific language support: JavaScript

Yes

Clear, concise documentation on setting up a Javascript project with Bitbucket pipelines: https://confluence.atlassian.com/bitbucket/javascript-node-js-with-bitbucket-pipelines-873891287.html

Yes

TravisCI is designed to be a simple way to integrate CI/CD in your workflow so it has a couple of features aimed at specific languages, such as Javascript, starting from pre-built containers (with node already installed, for example) all the way to automatically running specific platform commands (such as detecting a package.json in the root of the project and running npm test)

Integrations
1st party support for common tools (like Slack notifications, various VCS platforms, etc)

Yes

Large collection of available integrations: https://confluence.atlassian.com/bitbucket/apps-and-integrations-675189068.html

Yes

By default, TravisCI is built to work with GitHub. Additionally, there is strong support for 3rd party tools like Coveralls, BrowserStack, etc.

API
Custom integreation is available, via an API or otherwise, it's mentioned separately as it allows further customization than any of the Ecosystem/Integration options

Yes

Yes

Offers a feature-rich API that allows both reading data, as well as triggering or cancelling builds.

Auditing

Yes

N/A

Additional notes

There's some confusion regarding Bitbucket Pipelines and Bamboo, where they overlap and where not. Atlassian discontinued their Bamboo Cloud offering ~3 years ago, so at a high-level they are different products in that regard. What can be said about both is that they are top-tier tools for high-demand engineering teams, especially valuable as long as the other tools in the Atlassian suite are adopted (Bitbucket is a must for Bitbucket pipelines, being just one if it's features, but other tools like Jira are not to be dismissed either). It does seem like Bitbucket Pipelines is the more mature product of the two though.

Travis CI build matrix feature how to use it for CI parallelisation

Travis CI parallelism integration

Start using Knapsack Pro with Bitbucket Pipelines or Travis CI

Sign up and speed up your tests.

Get started free