Completed
Normalize per-key.
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Refining Systems Data without Losing Fidelity
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 Complex systems are hard to manage.
- 3 User experiences.
- 4 User experiences marbles.
- 5 without breaking the bank?
- 6 Three strategies for taming the spew.
- 7 Reduce. Reuse. Recycle.
- 8 Store less data.
- 9 Stop writing read-never data.
- 10 First, structure your data.
- 11 One event per transaction.
- 12 Often, trimming isn't enough.
- 13 Sample your data.
- 14 Statistics to the rescue!
- 15 Count 1/N events.
- 16 Count traces together.
- 17 Don't be afraid of sample rates.
- 18 Don't believe me? Ask a data scientist.
- 19 Aggregate data.
- 20 Aggregation destroys cardinality.
- 21 Temporal correlation is weak.
- 22 Math on quantiles is misleading.
- 23 Aggregation is a last resort.
- 24 How can sampling be cheap enough?
- 25 Systems scale with load.
- 26 Reconcile using the sample rate.
- 27 How can we save the relevant events?
- 28 Normalize per-key.
- 29 Different key, different probability.
- 30 Retain errors & slow queries.
- 31 Metrics and events can be friends!
