Thank You. 2021 Edition.
In the spirit of Thanksgiving, I am looking back at 2021 and thinking of all I have to be grateful for.
Confluent had a successful IPO and additional two successful quarters – and the wild stock market treated us rather well. I am now more financially comfortable than I expected to be, and perhaps for the first time in my life – I don’t need to lose sleep over financial concerns.
My team at Confluent has been excellent. In 2021 I’ve had five new engineers join the team and a new manager. I am grateful for those who joined – trusting Confluent to be a great place to work and trusting me with their career. I am also thankful for the engineers on the team for making everyone who joined feel welcome and for the fun and energetic team spirit.
I am deeply grateful to my husband, family, and friends for being supportive through the ups and downs of the year, my cats for being extra cuddly, and my body for continuing to function despite my age and benign negligence.
Our cloud vendors taught me, once again, that the cloud is just someone else’s computer. And cloud computing has its quirks and limitations and requires workarounds just like any other computer. We’ve learned that all cloud vendors have some limits to what they can provide and that we won’t necessarily know where they are until we hit them. We learned that when you hit the limit on the number of IPs, one cloud vendor will reuse an IP already in use by another service. We learned what happens when we modify LB configuration in massive K8s clusters with many services. We learned that cloud vendor maintenance could cause very noticeable latency spikes. Many important operational lessons, all learned the hard way.
Our customers taught us, once again, that they expect our managed cloud service to be far more magical than “running Kafka on someone else’s computer.” They expect things to “just work,” and when things don’t work, it is always due to an issue with the managed service. Cloud services are commonly delivered via thin REST APIs, so our customers are not all used to the idea that client or network configuration could cause problems. In addition, they are paying us to shield them from latency spikes, connectivity hiccups, and occasional stalls in the underlying cloud services. We made good progress in protecting our customers from hiccups and helping them use Kafka with minimal fuss, and there is also a way to go before we achieve our truly cloud-native vision and be as easy to use as S3.
My manager encouraged me to take on a project that I believed impossible – not due to technical challenges but organizational pushback. He taught me that I could make anything happen if I genuinely think that something is essential and needs to happen. I now demand more of myself and my team – if something is necessary, there must be an excellent reason why we are not doing it. “Someone else didn’t let me” won’t cut it. I am thankful for this lesson – I needed to learn it.
He also taught me how strategy, metrics, autonomy, and technical innovation interact. Too many technical leads shy away from setting measurable goals and driving to achieve them. But suppose you align an organization on a strategy and align on metrics that show how well you are executing this strategy. In that case, you gain a lot of autonomy in taking on new projects, as long as the improvement in those metrics matches the effort these projects require. Missing the goals set for these metrics serves to prioritize tasks that will help achieve these goals. I resisted taking on goals regarding the number of production incidents. In retrospect, these goals help me justify more automated testing and more investment in projects that will improve reliability. Important tasks would be harder to prioritize without these goals.
I also had great mentors this year. They taught me how to check that my skip-level reports are doing well, how to align with a lead engineer who works with my team but doesn’t report to me, new ways to resolve disagreements with thoughtful and well-meaning colleagues, and much more. I am grateful to everyone who had the patience and took the time to help me become a better technical leader.
And one last thing. Throughout my career, several people believed in me and trusted me beyond what I expected. I grew by retroactively earning this trust and making sure no one will ever regret trusting me. I am learning to do the same to others and help them grow by trusting them. I have incredibly ambitious 2022 plans. Those plans result from one awe-inspiring person appreciating what I do and trusting me to do much more, and from very few other impressive people trusting both of us beyond what looks reasonable.
I am incredibly grateful for everyone who believed in me – I wouldn’t be where I am today without this trust, and I will never take it for granted.