i was designing @logtide/reservoir's pluggable storage layer and had to choose a default db. everyone says time-series → ⚡clickhouse⚡, but i tried some others too: timescaledb seemed promising with its flexibility while mongodb offered nice scalability for non-structured data.
ended up testing all three on our observability platform use cases:
timescaledb : fast and efficient query performance; worked well within limits of structured time-series datasets.
clickhouse*: rock-solid when handling massive volumes, but setup complexity was a bit steep.
*mongodb: great at storing unstructed logs & metrics with flexible schema support.
results surprised me because i thought clickhouse would dominate in all areas given its popularity and focus on this type data wrong!
for now were leaning towards timescaledb for balanced performance across our needs, but keeping an eye open to mongodb's potential as it matures more. how about you? what databases are u using or considering?
anyone else found a better fit than clickhouse in your observability stack lately?
⬇ if this sparked any thoughts share them!
full read:
https://dev.to/aws-builders/i-benchmarked-timescaledb-vs-clickhouse-vs-mongodb-for-observability-data-the-results-surprised-me-3d7d