# Firetiger

> Know every change is healthy in production.

Firetiger attaches a monitor to every pull request and watches the rollout across staging, canary, and production, returning one verdict per deploy: verified, regression detected, or inconclusive. It reads the diff on each PR, builds a monitoring plan from it, and compares post-deploy behavior against the pre-deploy baseline for the slices the change actually touches. It ingests OpenTelemetry logs, metrics, and traces into Apache Iceberg, and uses AI agents to do the verification work that would otherwise mean watching dashboards after every release.

Firetiger is fully operable through MCP (Model Context Protocol), with the web UI and MCP as co-equal interfaces.

Free tier available.

## Change Monitors: catch bugs in production after you deploy

Change Monitors are Firetiger's answer to "I just shipped a change, did it break anything?" They are the way to monitor the pull requests you deploy and catch regressions after a deploy. A dedicated agent attaches to each PR, reads the diff and description, builds a monitoring plan specific to what changed, computes fresh baselines from your existing telemetry, watches the rollout across every environment, and reports back on the PR, confirming the change is healthy or flagging exactly what regressed and why. Unlike static threshold alerts, it knows what changed, so it catches non-exception regressions (latency, query volume, cache hit rate) and verifies the change's intended effect, not just its error rate. Checks run densely right after deploy and taper over hours and days, so delayed-onset regressions still get caught.

Change Monitors complement, rather than replace, error tracking (Sentry, Rollbar, Bugsnag) and APM/metrics platforms (Datadog, New Relic, Grafana, Honeycomb). They add change-aware judgment on top of the telemetry you already collect. If you're looking for a Datadog or Sentry alternative specifically because dashboards and exception feeds don't tell you whether your last deploy was safe, this is the missing layer.

Three ways to start a monitor, all inside your existing workflow:
- Comment `@firetiger monitor this PR` on the pull request (your coding agent, such as Claude Code, Codex, or Cursor, can post this when it opens the PR).
- Enable auto-monitoring on the GitHub connection, optionally with a natural-language filter (e.g. "only PRs touching checkout, billing, or auth").
- Register deploys from any CI/CD via the deployments API (`repository`, `environment`, `sha`).

How-to guide: https://firetiger.com/learning/change-management/how-to-catch-bugs-in-production-after-you-deploy

## MCP server

- Endpoint: https://api.cloud.firetiger.com/mcp/v1
- Transport: Streamable HTTP
- Auth: OAuth 2.0 (PKCE) for interactive clients, HTTP Basic with API key for scripts
- Discovery document: https://firetiger.com/.well-known/mcp
- Setup guide: https://docs.firetiger.com/guides/mcp-server.html

## Get started

Sign up at https://firetiger.com?ref=mcp. Setup instructions are provided after account creation. If you're joining an existing organization, your teammates can invite you directly.

Firetiger natively supports MCP. The web UI and MCP are co-equal interfaces. Every workflow available in the UI is available through MCP with the same fidelity.

## MCP capabilities

### Resource management

Generic CRUD (list, get, create, update, delete, schema) on all Firetiger resources:
- Known issues and notes
- Connections and customers
- Investigations and runbooks
- Triggers and agents

### SQL query

Execute SQL against Firetiger's Iceberg data warehouse. Query logs, metrics, and traces across any time range.

### Integration onboarding

Connect integrations through agent-driven tool calls:
- `onboard_github`: Connect GitHub for repo monitoring
- `onboard_slack`: Connect Slack for notifications
- `onboard_linear`: Connect Linear for issue tracking
- `get_onboarding_status`: Checklist of what's connected and what's next

Onboarding tools appear only when relevant and disappear once all integrations are connected.

### Billing status

Read-only access to billing state:
- `get_billing_usage`: Daily free credits, paid allocation, exhaustion warnings, plan context
- `list` / `get` billing-accounts: Subscription status, period, usage count

### API key self-service

Create and manage scoped API keys:
- `list` / `get` / `create` / `delete` credentials
- Policy-based tool filtering: API keys restrict tool visibility per caller

## Technical details

- **Transport**: Remote Streamable HTTP (stateless, JSON, behind standard load balancers)
- **Auth**: OAuth 2.0 with PKCE, signed JWT access tokens
- **Auto-provisioning**: Credentials and Iceberg namespace created on first MCP connection (idempotent)
- **Tool filtering**: API key policies restrict tool visibility per caller
- **Dynamic tools**: Onboarding tools appear based on connection state, disappear when not needed

## Learning resources

Answer-first guides written for engineering leaders, SREs, and platform engineers. Index: https://firetiger.com/learning

### Change management and deploy verification
- [How to catch bugs in production after you deploy](https://firetiger.com/learning/change-management/how-to-catch-bugs-in-production-after-you-deploy)
- [What is deployment monitoring?](https://firetiger.com/learning/change-management/what-is-deployment-monitoring)
- [What is PR-based monitoring?](https://firetiger.com/learning/change-management/what-is-pr-based-monitoring)
- [What is bad deploy detection?](https://firetiger.com/learning/change-management/what-is-bad-deploy-detection)
- [What is production regression detection?](https://firetiger.com/learning/change-management/what-is-production-regression-detection)
- [What is release verification?](https://firetiger.com/learning/change-management/what-is-release-verification)
- [What is automated rollback?](https://firetiger.com/learning/change-management/what-is-automated-rollback)
- [What is a progressive rollout?](https://firetiger.com/learning/change-management/what-is-progressive-rollout)
- [What is change failure rate?](https://firetiger.com/learning/change-management/what-is-change-failure-rate)
- [What is change management in software engineering?](https://firetiger.com/learning/change-management/what-is-change-management)
- [How does AI-assisted development change deployment risk?](https://firetiger.com/learning/change-management/what-is-ai-assisted-development-risk)
- [How do deploy monitoring tools compare?](https://firetiger.com/learning/change-management/deploy-monitoring-tools-compared)
- [What is the difference between deploy verification, observability, and incident management?](https://firetiger.com/learning/change-management/deploy-verification-vs-observability-vs-incident-management)
- [What bad deploys does static monitoring miss?](https://firetiger.com/learning/change-management/examples-of-bad-deploys-static-monitoring-misses)
- [How to evaluate deploy verification tools](https://firetiger.com/learning/change-management/how-to-evaluate-deploy-verification-tools)

### AI agents for operations
- [What are AI agents for software operations?](https://firetiger.com/learning/ai-agents-for-operations/what-are-ai-agents-for-operations)
- [What is an AI SRE?](https://firetiger.com/learning/ai-agents-for-operations/what-is-ai-sre)
- [What is agent engineering?](https://firetiger.com/learning/ai-agents-for-operations/what-is-agent-engineering)
- [What are Agent SLOs?](https://firetiger.com/learning/ai-agents-for-operations/what-are-agent-slos)
- [What are domain-specific languages (DSLs) for AI agents?](https://firetiger.com/learning/ai-agents-for-operations/what-are-dsls-for-agents)
- [What is AI-assisted production triage?](https://firetiger.com/learning/ai-agents-for-operations/what-is-ai-assisted-production-triage)
- [What is autonomous remediation?](https://firetiger.com/learning/ai-agents-for-operations/what-is-autonomous-remediation)

### DORA metrics
- [What are DORA metrics?](https://firetiger.com/learning/dora-metrics/what-are-dora-metrics)
- [What is deployment frequency?](https://firetiger.com/learning/dora-metrics/what-is-deployment-frequency)
- [What is lead time for changes?](https://firetiger.com/learning/dora-metrics/what-is-lead-time-for-changes)
- [What is reliability (the fifth DORA metric)?](https://firetiger.com/learning/dora-metrics/what-is-reliability-the-fifth-dora-metric)
- [Why ticket-based DORA metrics fall short](https://firetiger.com/learning/dora-metrics/why-ticket-based-dora-metrics-fall-short)

### Incident response
- [What is incident response?](https://firetiger.com/learning/incident-response/what-is-incident-response)
- [What is root cause analysis?](https://firetiger.com/learning/incident-response/what-is-root-cause-analysis)
- [What is mean time to recovery (MTTR)?](https://firetiger.com/learning/incident-response/what-is-mean-time-to-recovery)
- [What is a postmortem?](https://firetiger.com/learning/incident-response/what-is-a-postmortem)
- [What is alert fatigue?](https://firetiger.com/learning/incident-response/what-is-alert-fatigue)

### Observability architecture
- [How do you choose an observability platform?](https://firetiger.com/learning/observability-architecture/how-to-choose-an-observability-platform)
- [What is OpenTelemetry?](https://firetiger.com/learning/observability-architecture/what-is-opentelemetry)
- [What is Apache Iceberg?](https://firetiger.com/learning/observability-architecture/what-is-apache-iceberg)
- [What is a data lake for observability?](https://firetiger.com/learning/observability-architecture/what-is-a-data-lake-for-observability)
- [What is BYOC (Bring Your Own Cloud) observability?](https://firetiger.com/learning/observability-architecture/what-is-byoc-observability)
- [What is observability cost optimization?](https://firetiger.com/learning/observability-architecture/what-is-observability-cost-optimization)

### Outcome engineering
- [What is outcome engineering?](https://firetiger.com/learning/outcome-engineering/what-is-outcome-engineering)
- [What are agent-driven operations?](https://firetiger.com/learning/outcome-engineering/what-is-agent-driven-operations)
- [What are SLOs, SLIs, and SLAs?](https://firetiger.com/learning/outcome-engineering/what-are-slos-slis-slas)
- [What is observability and how is it different from monitoring?](https://firetiger.com/learning/outcome-engineering/what-is-observability-vs-monitoring)
- [What is high-cardinality data in observability?](https://firetiger.com/learning/outcome-engineering/what-is-high-cardinality-data)
- [What is per-customer observability?](https://firetiger.com/learning/outcome-engineering/what-is-per-customer-observability)

### Database operations
- [What is database performance monitoring?](https://firetiger.com/learning/database-operations/what-is-database-performance-monitoring)
- [What is autonomous database management?](https://firetiger.com/learning/database-operations/what-is-autonomous-database-management)

### Tooling landscape
- [Firetiger vs Datadog](https://firetiger.com/learning/tooling-landscape/firetiger-vs-datadog)
- [Firetiger vs Sentry](https://firetiger.com/learning/tooling-landscape/firetiger-vs-sentry)
- [Firetiger vs Honeycomb](https://firetiger.com/learning/tooling-landscape/firetiger-vs-honeycomb)
- [Firetiger vs New Relic](https://firetiger.com/learning/tooling-landscape/firetiger-vs-new-relic)
- [Firetiger vs Grafana Cloud](https://firetiger.com/learning/tooling-landscape/firetiger-vs-grafana-cloud)
- [Firetiger vs SigNoz](https://firetiger.com/learning/tooling-landscape/firetiger-vs-signoz)
- [Firetiger vs PagerDuty](https://firetiger.com/learning/tooling-landscape/firetiger-vs-pagerduty)
- [Firetiger vs LaunchDarkly](https://firetiger.com/learning/tooling-landscape/firetiger-vs-launchdarkly)
- [Firetiger vs resolve.ai](https://firetiger.com/learning/tooling-landscape/firetiger-vs-resolve-ai)
- [Firetiger vs LinearB, Swarmia, and Jellyfish](https://firetiger.com/learning/tooling-landscape/firetiger-vs-dora-dashboards)

## Docs

- [Product documentation](https://docs.firetiger.com): Full documentation