Has it really been another month already??? It doesn't seem possible, but here we are in November and we managed not to turn into a pumpkin 🎃. October was an excellent month at Pinpoint, where we worked to build on the momentum we carried through from September with our big launch. 

This blog is a candid retrospective on what we set out to accomplish in October.

The good news is we are a group of ambitious folks, but it was one of those months where we just didn't get to everything we planned. We had to adjust plans after the launch to solidify the product in stable and implement customer feedback. Regardless of not getting to everything on the roadmap, what the team accomplished really is game-changing. Here's a quick rundown of what was delivered:

  • Improvements and Fixes after launch - read the changelog here
  • Added Attachment support on the Issue Details screen
  • View Epic labels from Jira in Pinpoint
  • The Notebook 2.0 
  • New Integrations: CI/CD Data Models & GitLab Issues 
  • Real-time, bi-directional communication with Issue Management systems:
    • Updates to Issues including Status, Assignee, Comments, etc.
    • Ability to create issues directly in Pinpoint 
    • Auto-packed sprint plans and auto-assigned issues from Pinpoint are written to Issue Management Systems

Automated Work Planning

We have a goal to limit the time and steps involved in the entire process of building software from planning to deployment. We are doing this with machine learning, automation, and better integration into source systems like Jira to reduce jumping between tabs and the use of physical notebooks or scraps of paper. 

We had previously released our Work Planning features that auto-packs sprints, auto-assign issues, and identifies capacity challenges. Still, a significant roadblock was in our way to making this a really useful feature — the ability to have this work reflected back to the user's Issue Management system. 

With the release of our new Agent, we were finally able to close the loop on our Planning features. Now once a user has completed building their plan in Pinpoint, they can select "Write-back to..." to have the plan populate in the Issue Management system. This will save product owners and managers so much time from manually planning and inputting all the assignments. We also added "Create Sprint" to our Planning solution, so you never have to leave Pinpoint during that planning process. 

Notebook 2.0

Those physical notebooks and scraps of paper I previously mentioned? A thing of the past with our new and improved Notebook feature in Pinpoint. The Notebook has become something the team is extremely excited about and its evolution over the last 30 days is really impressive.

We have implemented a new editor (ProseMirror) as a part of our Notebook architecture and have added many unique capabilities along the way to make it smarter and more intuitive. The Notebook is now fully collaborative and live so teams and individuals can document, discuss, and get work done all in one simple and easy to use document. The Smart Blocks continue to evolve, and our use of Embed.ly brings lightning-fast embed speeds to any URL you may choose to bring into the Notebook.

Application Speed Improvements

We are fanatical about speed and bringing your data to you as fast as possible. As you are probably aware, this doesn't happen in one place in a SaaS-based application, and it doesn't happen overnight. When you are automating data science models, the challenge only grows (more on that here.) We have spent lots of time focusing on how to optimize for bringing data in under 400ms. This led us down a couple of paths.  

  1. First, we looked at our eventing infrastructure. We determined several inefficiencies here but were able to implement a solution that cut 1000s of messages down to a single message. This brought immediate improvements to speed.  
  2. The next area we looked at was how our app communicated with GraphQL and found some unnecessary latency and decided to switch to WebSockets to allow for lightning-fast communication between our web app and GraphQL.  
  3. The last update, which we mentioned in October, was discussed previously — our datamanager. Much of the data we present doesn't change often, so we knew there were many gains to be made caching.  

Combining these three updates has dramatically improved the speed, but we won't stop until we hit sub 400ms everywhere!

New Integrations

We had a pretty lofty goal of hitting four brand new integrations last month, and while we didn't hit them all, I'm excited about the ones we have added in Beta and what opportunities it will bring. We have our integration for Clubhouse, an Issue Management system ready to go and are looking for a few good beta candidates if you are interested. It was one of the most requested source systems from our users.  

We also were able to start work on supporting a new data source — CI/CD systems. The first CI/CD system we have integrated is CodeFresh. It's really early, so we will share more as it comes to light. We will continue to add to our integrations, but we will focus heavily on the Notebook, app speed, and re-imagining the traditional stand-up for November. 

So how do we feel that we did?

We feel really good about what we were able to accomplish in October, coming off the heels of a late September launch. While we didn't deliver everything we would have liked to, we made progress in each area we had intended to. Some of these things will carry forward with us into November, but that's part of iterating and improving daily and listening to our customers, using data to drive our decisions, and tracking our progress day in and day out.

Stay tuned for our November roadmap! 

- Scott 

Get the data science behind high-performance teams.