Status

System status

All systems operational

Postgres and Redis are both answering.

Postgres

operational

4ms

Redis

operational

3ms

Discord gateway

operational

15ms

Derived from the bot’s last published snapshot, which expires after two minutes.

Dispatch queue

clear

Processed

In flight

Backlog

Failed

Retrying Jobs waiting out a backoff before another attempt. Three attempts, then the job is kept for inspection.: 0
Kept for inspection Exhausted jobs still on the queue. Capped at 500, so this is a sample of recent failures rather than a total.: 10

Dashboard dispatches only. A prefix command typed in Discord never enters this queue, and the backlog reads 0 whenever the bot is consuming jobs as fast as they arrive — which is the normal state, not an idle one. Processed and failed count from the worker’s last restart, not from install.

Footprint

as cached by the bot

Servers

Members

10k

Channels

1.1k

Last hour

20 samples

Service	p50	p95	max	coverage
Postgres	12ms	33ms	100ms	20/20
Redis	3ms	15ms	31ms	20/20
Discord gateway	15ms	18ms	46ms	20/20

Coverage is how many samples in the window recorded a reading. It is not uptime — a missing sample can mean the collector did not run.

Gateway distribution

20 readings

14msbars past p95 are the tail46ms

Processes

since last restart

Gateway session: 1h 43m
Bot process: 1h 43m
Bot memory: 184.3 MiB
API process: 1h 43m

Restart ages, not availability. A deploy resets both, and neither figure can be read as an SLA.

Your connection

Round trip from this browser.

Snapshot age

The bot republishes stats on a two-minute expiry.

20s

Reporting coverage

last hour

Postgres40/40

Redis40/40

Discord gateway40/40

A neutral notch means no sample was recorded in that slice. That can mean the service was down or that the collector did not run, and this page cannot tell those apart — so it is not drawn as an outage, and none of these rows is an uptime figure.

Command activity

1 in 24h

-18h-12h-6hnow

Peak hour: 1. Columns are scaled against it, so the shape is relative to this window and not to any other day.

ranrefusederroredRefused covers a disabled command, a cooldown, or a missing permission.

Last 24 hours

100% ran

Outcome

Ran1100%

Missing permission00%

Rate limited00%

Disabled or gated00%

Errored00%

Source

Discord1100%

Dashboard00%

Execution timemedian 280ms, p95 280ms

Mean time inside the command engine, per hour. Hours with nothing in them break the line rather than being drawn as zero.

Latency trends cover the last hour, sampled every 15 seconds; command activity covers the last day, bucketed by hour. Neither is uptime — Karen does not publish an availability figure, so none is shown here.