← Back to the chapter · Previous detail

Technical detail · Data Retention and Downsampling

Chunk boundaries, compression, and holds

Retention shapeWhy it mattersRelease evidence
Aligned chunkThe cutoff falls on a chunk boundary, so a whole old chunk can be dropped.Chunk interval, cutoff timestamp, candidate chunk list, and dry-run output.
Straddling chunkThe cutoff lands inside a chunk, so the engine must wait or use a slower row path.Scheduler timezone, chunk boundary, and policy correction before automation.
Compressed warm dataOld chunks may be compressed before deletion, reducing warm storage cost while rollup and archive checks finish.Measured compression ratio, restore sample, and query replay on the compressed tier.
Held partitionAn investigation or compliance hold overrides the normal deletion unit.Machine-readable hold scope, owner, expiry review date, and release condition.

Boundary alignment is the trap. If chunks are one day long and the approved cutoff is midnight UTC, the engine can drop whole day chunks. If the job uses local time and the cutoff lands at noon UTC, the chunk straddles the boundary; the policy either waits until the whole chunk is expired or falls back to a slow partial-window path. Monitoring stores have the same idea with blocks: a Prometheus-style store removes old complete blocks, not arbitrary rows inside a block.

← Back to the chapter · Previous detail