By copying all logs into a centralized system, you no longer have to log into machines or VMs to manually
grep logs. Instead, you’re able to search across logs from many machines at once, applying filtering and aggregation pulled from Big Data tooling. This implies a reliable (and hopefully quick) method of forwarding log entries. This might be done via an agent or some other sort of sidecar process.
This allows teams to rely less on direct access to production compute, allows more people to perform analysis in parallel, and greatly improves their abilities due to the advanced features and user-friendly syntax of Big Data tooling (as opposed to custom logs processing shell scripts). Similarly, teams are less worried about losing a disk or a troublesome machine, knowing the logs are all safe.
Related Products: BigQuery, Cloud Logging
Related terms: TBC