Michael joined Airspace over seven years ago and has seen the company grow from a small startup to a large, 400 employee company. Airspace focuses on building a centralized platform for managing time-critical shipments in the logistics industry.
"After Aviator, we have a pretty high success rate for deployments. I'm not going to say it's 100%, but it's probably up there at like In the high 90s, 98, 99%. I like aviator because it's got good visibility. There's a UI, other developers can see into it, and support has been great."
The team was in a weird spot, as the need for a merge queue hadn’t been a priority. However, constantly restarting end-to-end tests was bogging down their CI/CD pipeline. It became apparent that a merge queue was necessary in order for frequent deployments to work.
"We used to have this heavy QA process, manual deploys. Originally, we would merge our PRs and then run a bunch of tests. If all the tests passed, then we would deploy. Now, we now run all the tests on the PR before we merge.
Requesting to merge created a bottleneck, and I knew it’d start building up as we scaled to more deployments.
We knew it’d be an issue to fix, and If it's not with this, it's with some other tool. We took a couple of surveys internally and some teams couldn’t wait to use Aviator.
One of the other big driving factors was to help our budget. We calculated that it’d cost $8 to run all of our tests on one PR because the codebase didn’t allow for checking code that’s changed."
"I came along, rolled up my sleeves and said, we got to get on CI/CD. So I set up some things and quickly realized like, wait a second. There's a need for a merge queue now.
We've made a couple of little config tweaks to Aviator here and there. What I really enjoy is it's almost like you set it up and forget about it.
We were in this weird spot too, cause we didn't actually need Aviator quite yet, but I was preparing my team before queuing issues became unsustainable."
Michael and team originally weren’t using CI/CD pipelines. After setting them up, the team quickly found the need for a merge queue to scale deployments. It took less than a day to create a proof of concept that gave the team context on how Aviator MergeQueue reacts with their primary repository.
Aviator’s support, visibility, and high success rate for deployments gave the team confidence to deploy frequently.
"After Aviator, we have a pretty high success rate for deployments. I'm not going to say it's 100%, but it's probably up there at like In the high 90s, 98, 99%. I like aviator because it's got good visibility. There's like a UI, other developers can see it into it. And support has been great. I can't stress how important enough that is from like a business perspective, especially when you got like a hundred things going on and you're trying to like get something accomplished. And then you get stuck and you don't hear from support for three or four days."
"If Aviator would not have been around, I would have just wrote something from scratch myself, but it wouldn't have been as nice as aviator. It would have been more just like fully tailored and custom to what we need to do."
Without Aviator, Michael would have been left to write a custom merge queue solution to manage PRs.