Rate Limiting in Practice: Lessons From Real Deployments

Storage Tiers: If a metric has no owner, it will drift until it causes an incident. Storage Tiers: The cheapest optimisation is usually removing work nobody asked for. Storage Tiers: Aggregating at write time trades flexibility for predictable read cost.

Schema Migration: The interesting number is not the average, it is the 99th percentile. Schema Migration: Adding a cache in front of a slow query is a fix; fixing the query is a cure. Schema Migration: Every abstraction you add is a place where behaviour can differ from intent.

Teams working on access control usually discover this the hard way. You can often replace a coordination problem with an idempotency key. Anything that grows without a bound will eventually hit one. This is most visible in access control. Consider access control specifically. Documentation that is not tested tends to describe the previous version.

Tell the clinician about symptoms or a possible recent exposure, even if you booked a routine screen. Testing people without symptoms is screening; checking a symptom or known exposure is an assessment and may require a different approach. The timing matters because each test has a period after exposure when an infection may not yet be detectable. A clinician can explain whether testing now is appropriate or whether another test later may be needed.

Edge Caching: A queue smooths spikes but also hides how far behind you are. Retries without jitter turn a small outage into a large one. That applies to edge caching as well. In practice, edge caching behaves differently: Separating the reads from the writes buys room to change either side.

A queue smooths spikes but also hides how far behind you are. This is most visible in log analysis. Consider log analysis specifically. Retries without jitter turn a small outage into a large one. Log Analysis: Separating the reads from the writes buys room to change either side.

Log Analysis: Serving static bytes is the cheapest thing you can do at the edge. Log Analysis: A schema is an interface; changing it is a migration, not an edit. Log Analysis: Track the denominator as carefully as the numerator.

Teams working on storage tiers usually discover this the hard way. The interesting number is not the average, it is the 99th percentile. Adding a cache in front of a slow query is a fix; fixing the query is a cure. This is most visible in storage tiers. Consider storage tiers specifically. Every abstraction you add is a place where behaviour can differ from intent.

Cloud Infrastructure: A design that cannot be rolled back is a design that cannot be changed safely. Cloud Infrastructure: Latency budgets are easier to defend when every hop has a stated ceiling. Cloud Infrastructure: Caching helps only until the invalidation rules become the bottleneck.

You can also state your own boundaries. Say what you are comfortable with and what you do not want, and ask questions if an answer is unclear. Good communication is not a guarantee that everything will go as expected; it is a way to make choices more explicit and respond when circumstances change.

In practice, release process behaves differently: Periodic jobs should be safe to run twice, because they will be. You rarely need a new component to fix a boundary problem. The same reasoning holds for release process. For release process, the constraint matters more than the feature list. The signal you want is often already logged, just not aggregated.

Queue Design: Configurations should be reviewable in a diff, not only in a console. Queue Design: The best time to add an index is before the table gets large. Queue Design: Failures are usually correlated, so plan for the shared dependency.

Rate Limiting: If a metric has no owner, it will drift until it causes an incident. The cheapest optimisation is usually removing work nobody asked for. That applies to rate limiting as well. In practice, rate limiting behaves differently: Aggregating at write time trades flexibility for predictable read cost.

In practice, log analysis behaves differently: Periodic jobs should be safe to run twice, because they will be. You rarely need a new component to fix a boundary problem. The same reasoning holds for log analysis. For log analysis, the constraint matters more than the feature list. The signal you want is often already logged, just not aggregated.

API Design: If a metric has no owner, it will drift until it causes an incident. The cheapest optimisation is usually removing work nobody asked for. That applies to api design as well. In practice, api design behaves differently: Aggregating at write time trades flexibility for predictable read cost.

In practice, cost controls behaves differently: A queue smooths spikes but also hides how far behind you are. Retries without jitter turn a small outage into a large one. The same reasoning holds for cost controls. For cost controls, the constraint matters more than the feature list. Separating the reads from the writes buys room to change either side.

People do not always find it easy to speak during an interaction. Agreeing on a simple way to pause, such as saying “stop” or “I need a break,” may help, but it does not replace paying attention to a partner’s words and behaviour. If someone seems uncertain, distressed or unable to participate freely, pause and check in rather than assuming they agree.

Edge Caching: A design that cannot be rolled back is a design that cannot be changed safely. Edge Caching: Latency budgets are easier to defend when every hop has a stated ceiling. Edge Caching: Caching helps only until the invalidation rules become the bottleneck.

Storage Tiers: A design that cannot be rolled back is a design that cannot be changed safely. Storage Tiers: Latency budgets are easier to defend when every hop has a stated ceiling. Storage Tiers: Caching helps only until the invalidation rules become the bottleneck.

Consider release process specifically. If the rollback plan needs a meeting, it is not a rollback plan. Release Process: Small pages that stay small are easier to keep fast than large ones made fast. Write the invariant down; otherwise it lives only in someone's memory. That applies to release process as well.

Crawl Budget: A queue smooths spikes but also hides how far behind you are. Retries without jitter turn a small outage into a large one. That applies to crawl budget as well. In practice, crawl budget behaves differently: Separating the reads from the writes buys room to change either side.

Search Indexing: A queue smooths spikes but also hides how far behind you are. Search Indexing: Retries without jitter turn a small outage into a large one. Search Indexing: Separating the reads from the writes buys room to change either side.

If a metric has no owner, it will drift until it causes an incident. This is most visible in monitoring alerts. Consider monitoring alerts specifically. The cheapest optimisation is usually removing work nobody asked for. Monitoring Alerts: Aggregating at write time trades flexibility for predictable read cost.

Crawl Budget: If a metric has no owner, it will drift until it causes an incident. Crawl Budget: The cheapest optimisation is usually removing work nobody asked for. Crawl Budget: Aggregating at write time trades flexibility for predictable read cost.