Ask HN: Was hired to improve company's devops, founder won't listen to my ideas
So after two weeks, I got the approval to create a staging branch. I wrote up a huge document on the staging workflow, covering edge cases such as hotfixes and rollbacks. I put time on people's calendars and walked them through it including the founder. Yet he still continues to resist the change and pushes directly to master, ignores the new processes I put in place, and generally just does what he wants. He also keeps harping that I need to "earn his trust." I find this baffling - I mean, I was hired for a reason right? Why would they hire me if they didn't trust me?
I am starting to get frustrated, especially since the founder has pushed breaking changes directly to master and then I got flack since I am now the "systems" guy and am responsible for the health of the app. I'm trying to think of the best way to approach this diplomatically. Do any HNers have some pointers or advice to give? I've worked with difficult clients before but I've never faced such a huge lack of trust.
7 comments
[ 3.2 ms ] story [ 28.3 ms ] threadIf logic is breaking, look to encourage units tests on new code, and back port unit test coverage as time goes on.
The idea is to allow everyone to suggest and implement technical improvement by providing a process to do so. For each idea, the owner fills a spec to describe it. The main ideas are: - Be transparent on what we want to do and our issues, the doc should be public and comments-friendly - Follow an advice process by proactively asking people to improve or challenge the idea - Always have by a trial with time boundaries and goal
A template could look like this:
- Owner: Name
- Title: to easily speak about it
- Reviewers: people who should definitely have a look
- Summary
- Why: Describe what are the issues with the current process, the bugs or similar.
- What is the change
- How to implement the change
- Duration of the trial
- What should we've achieved at the end of the trial
In your situation, it might be something like:
- WAthrowaway
- Staging env
- Reviewers: Founder, Senior Eng #1...
- Setup a staging environment
- We encountered critical issue #1 on October 1st, issue #2 on October 2nd
- A staging environment would allow to catch this bugs before it is on production
- 1. Create a staging env 2. Create a staging branch 3. block push on master 4. PR staging / master
- 1 month
- Critical issues happen on staging and not on master
PS: It's inspired by http://www.reinventingorganizationswiki.com/Decision_Making