- #Masstransit azure function code#
- #Masstransit azure function series#
- #Masstransit azure function windows#
Even for my initial design I considered using Azure Functions to build it but decided against it due to the constraint of needing to run on-premises. This also meant that the pull request bot had to move along with it to the cloud, so this was an ideal opportunity to rethink its design.
Looking at the cloudĪbout a year later we made the decision to migrate our on-premises TFS onto Azure DevOps hosted in the cloud. Maintaining it was mostly down to me though and due to the complexity it wasn’t easy for my co-workers to help me with that. From a usage perspective it did much of what I had initially hoped to deliver so I was quite happy with it.
#Masstransit azure function windows#
NET Framework applications running on IIS on Windows Server, so this was quite a departure from that.īut the good news was that it was working and it was working well. Most of my co-workers were used to developing. In fact, that was exactly the feedback I got from some of my co-workers when I presented the slide shown above at an internal talk. If you’re reading this and thinking “Wow, that is quite complex”, you are absolutely right. And of course I needed to do some logging as well, so I used Serilog to write semantic logs to Elastic Stack. For keeping the state of what comments had already been posted I ended up using Marten to turn PostgreSQL into a document database. To avoid locking up our TFS server I figured it might be good to introduce some queuing, so I picked MassTransit and RabbitMQ to solve that bit. I initially approached this as a hobby project of sorts, so I figured it was a good time to try out some new technology for this project.NET Core 2.0 just came out back then and Docker was all the rage, so I went with that. I didn’t want to flood the developers with a sea of comments, so I needed to keep some state as to what comments had already been posted.Cloud was a bit of a no-go for us at the time, so it had to run on-premises.
#Masstransit azure function series#
This is the first post in a series and I figured it would be a great story to tell as part of the Applied Cloud Stories initiative.įor my first try I had a couple of constraints and challenges that I needed to tackle: I wanted to develop something that could be easily extended to perform whatever automated checks we could think of and then post comments in the pull request to trigger developers. And there the idea of a pull request bot was born. So I suggested to do some automation around these things, since they were laborious and dull which made them an ideal candidate for automation. This usually caused delays and rework later on in the development process. However, we found that some things that would need to happen for every pull request weren’t being done (or not done properly) which weren’t always caught by reviewers either.
#Masstransit azure function code#
For example, we introduced pull requests as a mandatory step in the development flow in order to increase awareness and code quality. At the time we had recently switched from TFS Version Control (TFVC) to Git, which obviously introduced quite a few changes in the way we worked. About 2,5 years ago I was talking to my fellow team members about some of their challenges.
0 kommentar(er)
