Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.tracelit.io/llms.txt

Use this file to discover all available pages before exploring further.

Tracelit automatically opens an incident when something in your backend goes wrong — a spike in errors, a new exception type, an anomalous metric, or an SLO burn. No manual alerting rules are required to get started.

How incidents are detected

Tracelit continuously processes your telemetry signals and opens an incident when it sees:

Error spike

A sudden increase in error rate for a service or endpoint above the baseline.

New error type

An exception fingerprint that has never been seen before in this service.

Metric anomaly

A metric deviates significantly from its historical pattern (latency, queue depth, etc.).

SLO burn

Your error budget is being consumed faster than the allowed rate.

Incident lifecycle

Every incident moves through a defined lifecycle:
1

Open

Tracelit detects an anomaly and creates the incident with severity, trigger payload, and an AI-generated summary.
2

Acknowledged

A team member has seen it and is investigating. Notifications stop escalating.
3

Resolved

The issue is fixed. Tracelit closes the incident and records the resolution time.
You can also ignore an incident if it is a known non-issue — Tracelit will not re-open it for the same fingerprint.

What each incident includes

FieldDescription
Severitycritical, high, medium, or low — set by signal type and volume
AI summaryOne-sentence description of what went wrong and likely cause
Trigger payloadThe raw signal (error, metric, trace) that opened the incident
Linked traceThe trace closest to the incident trigger, with full span tree
TimelineChronological comments and status changes
Resolution linkDirect link to the AI-generated fix PR, if one has been created

Notifications

Tracelit can notify your team the moment an incident opens. Configure destinations in Settings → Integrations.

Slack

Incidents posted to a dedicated channel with severity, summary, and quick-action buttons (Acknowledge / View trace / Create fix).

Microsoft Teams

Same rich incident cards delivered directly to your Teams channel.

Alert rules

In addition to automatic detection, you can define custom alert rules on any signal type:
SignalExample rule
errorAlert when payments-api returns more than 10 errors per minute
metricAlert when job_queue.depth exceeds 1,000 for 5 minutes
logAlert when a log line matching "FATAL" appears in auth-service
traceAlert when p99 latency on POST /orders exceeds 2,000 ms
Configure rules in your service settings under Alerts.

From incident to fix

When your service has a GitHub repository connected, Tracelit can automatically generate a fix and open a pull request directly from an incident.

GitHub PR integration

Connect your repo and let Tracelit go from incident to ready-to-review PR in minutes.