Reference/Limits and defaults

Limits and defaults

The numbers that decide behaviour, collected in one place.

Every number here is a constant in the code, not a setting, unless a variable is named. Where a default is configurable, the environment reference says how.

Alerting

	Value
Notification retries	3 (4 attempts in total)
Retry backoff	1 min, 5 min, 30 min
Retry claim lease	5 min
Slack section text	3000 characters (enforced at the transport)
Slack field text	2000 characters, 10 fields
Slack blocks per attachment	50
Telegram message	4096 characters

See Notification delivery.

SQL alerts

	Value
Query timeout	2 minutes
Rows evaluated per run	2000
Rollup lines per message	25
Rollup body	2400 characters
Worker tick	1 minute
Clock-alignment offset	0–1439 minutes

See SQL alerts.

Agent

	Value
Version	`0.11.0`
Health document read cap	2 MB
Health flush interval	15 s
Health batch	30
Transport	Outbound websocket only

See Agent reference.

Cloudflare

	Value
Load-balancing adaptive dataset retention	~3 days on this plan
HTTP adaptive dataset retention	Much longer
Region→pool query row limit	10 000
Region→pool chunk window	26 hours
Region→pool capture bucket	5 minutes
Routing matrix rows returned	500

Important

The three-day wall is why AccelerUp records the region→pool split itself. A seven-day routing matrix is only a measurement because of that capture — see Load balancing.

API

	Default	Variable
API rate limit	600/min	`RATE_LIMIT_API`
Auth rate limit	30/min	`RATE_LIMIT_PUBLIC`
OAuth rate limit	120/min	`RATE_LIMIT_OAUTH`
Session TTL	24 h	`JWT_TTL_HOURS`
DB pool	25	`DB_MAX_CONNS`

Ingest

	Default	Variable
Events per request	100	`WEB_INGEST_MAX_EVENTS`
Body size	256 KB	`WEB_INGEST_MAX_BODY_BYTES`
Rate per IP	240/min	`WEB_INGEST_RATE_PER_MIN`

Retention

	Default	Variable
Metric tables	90 days	`RETENTION_METRICS_DAYS`
Telemetry tables	90 days	`RETENTION_TELEMETRY_DAYS`
Per-organisation policies	—	Data retention

Part of Reference — The exact numbers, names and limits, generated from the code.