Infrastructure Team

People

Team members

Ellie Huxtable
Team lead
Full Stack Engineer
Luke Harries
Head of Product
Daniel Esneider
SRE Engineer

Mission

Make deploying, scaling, and managing PostHog easy, fast, and reliable.

Q1 2023 Goals

Objective: Reduce TOIL
- KR: Migrate PostgreSQL out of Heroku and into the correct AWS account
  - Why? Reduce cost and reduce risk of us maxing out the instance
- KR: Consolidate infrastructure into the correct AWS accounts (specifically US)
  - Why? This will reduce the toil and AWS expenses
Objective: Support other teams and enable other teams to self-serve
- KR: Most of the pipeline team has written/contributed to their own alerting system
  - Once the system is set up well, would be good to get the other teams doing this too
  - Why? Enable the pipeline team to self-serve their issues
- Metric to track: % of non-infra hero time supporting other teams
Objective: Keep infra spend flat at $120k/month
- Such as Guido’s list
- Will likely decrease cost first then increase with usage
- Why? Improving our gross profit margin

Roadmap

3-year

All infrastructure is managed as code
Cloud is global
Best in class security and privacy compliance
Scale beyond 1 Trillion events / month
Support non-k8s Deploys 🤖
Painless developer/contributor experience

6 month

No Heroku
5 Billion / month events
SOC 2 Ready
All infra is managed as code on prod / staging + EU Ready 🎈
Rock solid, reliable infrastructure with regular production load testing to inform scaling plans

Slack channel

#team-infrastructure

How we work?

Guidelines

We work as teams on one goal/project - not having a single person alone working on a goal
The board should be our source of truth
We document what we do to share context internally
We finish what we start, or we don't start it at all
We continually prioritize
We prioritize unblocking others
We have an agenda and follow up on actions from our meetings
Be frugal

Standups

We have a Platform wide standup every Monday, Wednesday, and Friday. Standups are an opportunity for us to discuss what we are working on, feedback, and topics we may want other people's opinions on. It is also an important forum to announce that you are blocked or ask for help. Everyone should try to make standups but feel free to drop off if what is being talked about isn't relevant or valuable to you.

Engineering Planning

We plan our work using a two week sprint with sprint planning and retro meetings on the wednesday before the start of the next print. We primarily use the Platform project board to communicate what we are working on for the sprint, what is blocked, in review, done, and what we are planning on doing next.

Sprint Planning

Sprint planning happens every other week the Wednesday before the start of the new sprint and is PostHog engineering wide. We first break out into breakout rooms for Infrastructure and Ingestion and in each we determine what the goals will be for the upcoming sprint. After that we join back into the engineering wide sprint planning we pitch the goals to the entire engineering team, looking for feedback. This is a great opportunity for other teams within the company to raise concerns about things they may be blocked on.

Sub-teams are fluid so members may change from sprint to sprint.

Retro & Team Planning

After the Eng Sprint Planning meeting the Infrastructure and Ingestion teams will meet up to retro the previous sprint. This is also where we game plan the next sprint in terms of what tasks need to be done to accomplish to goals set in the Engineering Sprint Planning.

While planning we make sure that the teams that we have settled on for Ingestion and Infrastructure have more than one person working on a goal or project in the same timezone. We want to reduce the number of lone wolfs and encourage people to work together and spread context. There are a few benefits to this including shared context, quicker PR approvals, easy rubber ducking, and more trust and camaraderie on the team. 🌞

Project Boards

We use Github Project boards to organize what work needs to get done for a certain project. During a sprint we may not get an entire project done, but we should set our goals relative to milestones measurable in the project boards.

For example if there is a project that is to re-partition 100 tables, goals set for a sprint could look like:

Migrate 50 of the 100 tables
Migration framework is production ready
We have migrated 10% of customers with all 100 tables

The projects can be viewed as epics if that is what you are used to.

Team Kanban Board

We use a Kanban style board for each team (Infrastructure and Ingestion) to show what we are working on, planning on working on, blocked on, what is in review, and what is done. This provides context on what the operational priorities are for the sprint and what work people can pick up if they have a few extra cycles or are looking for what the most impactful task is at any one moment for them and their sub-team. It's up to the sub-teams to decide what tasks are on deck for each sprint based on the goals that were set for the sprint. We try to keep this as up to date as possible and assign ourselves as owners so that there isn't duplicate work done on the same task and if there is a question about a task we know who to ask.

We also tag the tasks that we set aside for the sprint with the sprint number/name so that we can filter out what is on the board for a quick view of how we are making progress against the sprint goals.

The board also acts as a source of truth for other teams to have a quick check in on the progress of tasks for the sprint, especially if they will be the primary consumers of the product of the task. This works within the team as well.