You can scale things other than just servers. I used my engineering blog search engine FullJoin.io to find blog posts on how companies scaled their processes, data centers, and yes - servers.
1. Airbnb - Used caching to buy time to fix their architecture issues
Key take away: If you hit a wall with the load that your databases can handle, lots of caching can give you breathing room until you can rearchitect your app for increased traffic.
2. The New York Times - Tips for scaling an engineering team
Interesting tidbit: They used pair programming as a way of transmitting the cultural values of their team to new hires.
3. Coinbase - Scaling the deployment process as they added more services and engineers
Interesting tidbit: They went from deploying 128 times per year to 580. Investing in the deployment process made it safer and less scary to do for new engineers.
4. Facebook - Created a second data center on the east coast to lower latency
Key takeway: When you're at Facebook scale, you can't just spin up a few extra AWS servers. It's more cost effective to create your own data center.
5. Dropbox - Migrating from S3 to their own exabyte sized data center (series)
Key takeway: Similar to Facebook, they had the scale to need their own datacenter, but they continued using AWS where it made sense.
6. Jet - Different ways of scaling microservices
Key takeway: There are different ways at varying levels of the stack to tune microservice performance.
7. Intercom - Interview with VP of engineering on scaling the engineering team
Interesting tidbit: Intercom hires for potential, since they're growing so fast and they need to hire people who can adapt to future challenges.
8. Cockroach Labs - How the hiring process scaled the company from 56 to 117
Interesting tidbit: To avoid being biased by someone's background, the engineers at Cockroach Labs don't look at resumes.
9. LinkedIn - Sending 2.75 gigabytes of data over Kafka
Interesting tidbit: Having core Kafka committers on the team has let LinkedIn push their Kafka infrastructure to the limit, while stilling keeping things maintainable.
10. Pinterest - Autoscaling servers during peak hours
Key takeaway: Sometimes something from your cloud platform doesn't work exactly how you need it to, so you need to supplement it with something built in-house.
11. Slack - Adding Kafka to their Redis job queue system to prevent downtime
Key takeaway: Sometimes when you hit scaling problems, rather than rearchitect your whole system, you can add a component that stabilizes the system.