Skip to content

performance.md

Latest commit

 

History

History
30 lines (22 loc) · 1.07 KB

performance.md

File metadata and controls

30 lines (22 loc) · 1.07 KB

Performance

Tested on the following environment with a pre-generated database of ~43200000 checks data (1000 dummy endpoints, 30 days of history, interval of 60 seconds) with total of 3.4G DB Storage.

# To run with constrants
systemd-run --user --scope \
    --unit=tinypulse-stress \
    -p CPUQuota=100% \
    -p MemoryMax=1G \
    -p MemorySwapMax=0 \
    ./tinypulse --addr 3000

# To read
systemd-cgtop -p /user.slice/user-$(id -u).slice/$(systemctl show --property ControlGroup --value tinypulse-stress.scope)

A quick chart.

Also. hydration only takes a couple seconds for this database size.

time=2026-04-21T16:12:25.614+08:00 level=INFO msg="hydrating state manager from database..."
time=2026-04-21T16:12:30.755+08:00 level=INFO msg="state manager hydrated successfully" endpoints=1000
time=2026-04-21T16:12:30.758+08:00 level=INFO msg="server starting" addr=:3000

On a Mac M1 (development device) - with the datapoints as above - had about ~175 doing rest. During hydration or aggregation this might spike to 200MB.