Over the last year, I was handed a bunch of projects in various states. One of the first things I do when getting a new project that I'm suddenly responsible for is to audit the project. That helps me figure out what I'm looking at and what I need to do with it next.
This blog post covers my process for auditing projects I'm suddenly the proud owner of.
What do I mean by "audit"?
When I was younger, I'd wander around a project and figure it out as I went along. That takes a long time during which I don't really know what I'm doing, I don't feel good about it, and I'm learning too many things from dealing with nasty surprises. I neither enjoy that nor does it make me look good.
These days, I take a methodical approach to picking up a project. I spend a week or so working through a set of questions. I find this approach makes for a good survey of the project and the problem domain it exists in. Further, it surfaces the grime that I should clean up thus avoiding nasty surprises.
The primary audience for the audit is me--I'm using the process as a way to come up to speed on something. I'm a software engineer so I'm focused on software development and maintenance aspects of the project. I'm concerned about security and data policies, uptime and reliability, impact on stake-holders, costs, budgets, infrastructure complexity, and ongoing maintenance work.
The next audience is my manager and co-workers and whoever else I might pass this project on to or might be interested.
I use the word "audit" to cover that period of time just after I've been given a project where I'm trying to figure out what it's all about. I want to know what the project is all about, what state is it in, how do I maintain it, who else is involved, and what do I need to change to make it maintainable.
I break it up into four parts:
What is it?
What are the project details?
How is it maintained?
What is the current status?
I start a document and work through questions and answers. For questions I don't know the answer to, I'll add a note to look into it later and move on. I keep iterating over the document until I've fleshed it out enough that I feel like I have a good grasp of the project.
Let's walk through the parts.
Part 1: What is it?
Why does this project exist?
What does this project do?
What is the context in which it exists?
Who has worked on this project in the past?
Who is currently working on this project?
For whom does this project exist?
Are there other projects that depend on it?
What are they?
Are there possible future users that this project is working towards?
Part 2: What are the project details?
What are the URLs to project website, documentation, license, runbook, and wiki?
What are the URLs to project repository, issue tracker, road-map, and development planning?
What are the URLs to IRC channels, Slack, Discourse, Telegram, mailing lists, Matrix, and other forums that the project uses?
What are the URLs to CI, metrics dashboard, site status, Pingdom, logs, and anything else for observing the health and status of the project?
What are the major components, services, storage systems, queues, etc for the project?
What data does the project use and how does it flow through the system?
What languages, versions, and runtimes are used?
What infrastructure is used? How is it defined?
Is there a system for authentication/authorization? How does it work? Who is responsible for the systems involved?
Part 3: How is it maintained?
What version control system is used?
What version control processes are used?
Is there a master repository? If so, where is it hosted?
What are the requirements for the project? * Uptime requirements? * Browser support matrix? * API compatibility requirements? * Et cetera
What is the quality assurance story for the project and how does it ensure the requirements?
Where are the test suites? What do they test?
When are tests run? What's tested in CI? What's tested by hand?
Which linters are used? What do they lint? When is linting run?
What processes ensure dependencies are up-to-date?
What processes watch for security issues in dependencies?
How is the project deployed/released? Is it written down somewhere? Who needs to be involved to do it?
How often is it deployed/released?
What gets logged?
What is being measured to determine whether requirements are met?
What is being measured to determine health of the system?
Where are unhandled errors captured? What monitors them?
How long are logs kept for? Who has access to the logs? Are logs archived somewhere? How long is that kept for?
How long are metrics kept for? Who has access to metrics? Are metrics archived somewhere? How long is that kept for?
What personally identifiable information is captured by the project? Where is it stored? How long is it stored for?
Part 4: What is the current status?
What periodic maintenance is required for the project? When was it last performed?
When was the last deployment/release? What has changed since then? Is the project deployable/releasable now?
Are dependencies up-to-date? Are any of the dependencies in use obsolete, abandoned, or deprecated?
Are there things that are important or required, but not covered by tests?
When was the last security review done on the project? What was the outcome? Are there any security issues currently? Should it have another security review?
Are there any in-progress projects? Technical debt cleanup? Migrations? What state are they in? What's the urgency? What's the next steps?
What urgent things need to be done on this project?
I work through the questions that are pertinent to the thing I'm auditing and answers and often that surfaces other questions that need answers. I keep iterating over that until I end up with two things:
An audit document.
I can use this to do my work as I become more familiar with the project. I can show this to my manager and coworkers so they're familiar with the project. I use this to improve the runbook, FAQ, README, and other project documentation.
A prioritized list of things to do next.
Some of the things will already be in an issue tracker. Some of the things will not be. I write up issues for all the things that should get done. I prioritize them and hit the ground running.
I have one co-worker. In Summer of 2019, we managed to pick up 5 additional projects in various condition, audit them, and get them into a maintainable state. We couldn't have done that without being methodical about picking up projects.
If you find this helpful, let me know!