How many people use the app, how often, and at which
hours — all via anonymous aggregated counts (without knowing who anyone is).
Unique users per day (new vs. returning)
Activity by hour of day
Activity by day of week
Searches per active user (per day)
How long the user waits for the full response and where that time is spent.
Response time distribution
Latency over time
Average latency by subsystem (ms)
By subsystem
Subsystem
Samples
Average
p50
p95
Health of the third parties we depend on. High latency or errors mean adjust cache / reduce queries.
Provider
Calls
Error rate
Avg latency
p95
Last failure
Third-party credentials are encrypted on the server
and never appear in files, logs, or on this screen. Enter a value to set or
rotate it; we only show a short “fingerprint” so you can confirm which
value is active.