Our team was launching a new feature and one of the backend engineers got pulled onto another project right before the deadline. I stepped in and picked up his remaining tasks — writing the integration tests he hadn't finished and helping review his PRs. My own work was already done so I had the bandwidth. We shipped on time and my manager gave me a shoutout in the team meeting. It felt like the right thing to do — I didn't want the team to miss the deadline because of a resourcing gap.
Three weeks before a major launch, I noticed our error-rate monitoring covered only our own service — nothing downstream. I flagged it to my manager, but the on-call rotation owner said it was out of scope for my team. I disagreed and built a lightweight cross-service alerting dashboard anyway, spending evenings over two weeks. On launch day, it caught a latency spike in a dependent service within four minutes — our previous mean-time-to-detect was over forty minutes. I documented the monitoring gap as a post-launch action item and proposed it as a standard for future launches. That mechanism was adopted across three other teams.