Configuration reference

config/telemetry-insights.php. Publish it with php artisan vendor:publish --tag=telemetry-insights-config.

Master switch

Key	Env	Default
`enabled`	`TELEMETRY_INSIGHTS_ENABLED`	`true`	Off registers no commands, schedule or pages

Key	Env	Default
`scope.service`	`TELEMETRY_INSIGHTS_SERVICE`	all	Limit the scheduled work to one service
`scope.environment`	`TELEMETRY_INSIGHTS_ENVIRONMENT`	all	Limit it to one environment
`scope.window`	`TELEMETRY_INSIGHTS_WINDOW`	`15m`	How far back each pass reads

Keep the window comfortably longer than the scan interval, so a slow ingest never drops an error between passes.

Key	Env	Default
`correlation.burst_seconds`	`TELEMETRY_INSIGHTS_BURST_SECONDS`	`120`	How close two groups must start to be one event
`correlation.min_groups`	`TELEMETRY_INSIGHTS_MIN_GROUPS`	`3`	Below this a burst is not an incident
`correlation.shared_ratio`	`TELEMETRY_INSIGHTS_SHARED_RATIO`	`0.6`	Share of the burst that must touch the failing dependency
`correlation.max_trace_probes`	`TELEMETRY_INSIGHTS_MAX_PROBES`	`20`	Trace fetches per pass — this bounds the cost
`correlation.sample_limit`	`TELEMETRY_INSIGHTS_SAMPLE_LIMIT`	`500`	Occurrences read per pass
`correlation.change_window_minutes`	`TELEMETRY_INSIGHTS_CHANGE_WINDOW`	`120`	How far back to look for a deploy to blame
`correlation.signal_pad_seconds`	—	`300`	Padding around onset when reading host signals

Key	Env	Default
`digest.limit`	`TELEMETRY_INSIGHTS_DIGEST_LIMIT`	`5`	Findings per kind
`digest.slow_route_ms`	`TELEMETRY_INSIGHTS_SLOW_ROUTE_MS`	`1000`	p95 above this is a finding
`digest.slow_query_ms`	`TELEMETRY_INSIGHTS_SLOW_QUERY_MS`	`100`	Average above this is a slow query
`digest.repeated_query_calls`	`TELEMETRY_INSIGHTS_REPEATED_QUERY_CALLS`	`100`	Calls above this is a repeated query

Key	Env	Default
`notify.default`	—	`['log']`	Channels used when a rule names none
`notify.channels`	—	log, slack, webhook	Name → class map
`notify.log.level`	`TELEMETRY_INSIGHTS_LOG_LEVEL`	`warning`
`notify.slack.webhook_url`	`TELEMETRY_INSIGHTS_SLACK_WEBHOOK`	—
`notify.webhook.url`	`TELEMETRY_INSIGHTS_WEBHOOK_URL`	—

Key	Env	Default
`announce.new_issues`	`TELEMETRY_INSIGHTS_ANNOUNCE_NEW`	`true`
`announce.regressions`	`TELEMETRY_INSIGHTS_ANNOUNCE_REGRESSIONS`	`true`
`announce.incidents`	`TELEMETRY_INSIGHTS_ANNOUNCE_INCIDENTS`	`true`

Off means recorded silently, not ignored.

Key	Env	Default
`schedule.scan`	`TELEMETRY_INSIGHTS_SCAN_CRON`	`/5 * * *`
`schedule.alerts`	`TELEMETRY_INSIGHTS_ALERTS_CRON`	`/5 * * *`
`schedule.digest`	`TELEMETRY_INSIGHTS_DIGEST_CRON`	`0 8 * * 1`

Set one to an empty string to schedule that job yourself.

Renaming a table here is enough; the migration reads the same config.