Introduction
An open-source status page you deploy into your own AWS account.
Galena checks your services every minute from three AWS regions, decides when something is really down, and publishes a static status page that keeps serving when everything else is not. It runs in your own AWS account, alongside trigger.dev for background work.
What it does
- Checks every HTTP monitor once a minute from three probe regions, with DNS, connect, TLS and time-to-first-byte timings for each check.
- Decides a monitor is down only when enough regions agree, and never on a single failed check. A region whose own network looks broken leaves the vote, and a monitor that keeps flipping is marked flapping instead of paging people every minute.
- Publishes a status page as static files on S3 and CloudFront, in a region of its own, with a copy in a second region. Visitors never reach the API or the database, so the page stays up when they are down.
- Tells people by email (double opt-in, one-click unsubscribe), Slack and signed webhooks when an incident or maintenance window is posted.
- Lets your team run incidents and maintenance from a dashboard with roles and two-factor sign-in.
How it fits together
Galena is three paths that fail independently.
| Path | What runs | What it does |
|---|---|---|
| Hot path | Probe Lambdas in three regions, an SQS FIFO queue, an evaluator Lambda, DynamoDB | Checks every monitor every minute and works out each monitor's state |
| Cold path | The API on Lambda, Aurora Serverless v2, trigger.dev tasks | Incidents, maintenance, notifications, and publishing the page |
| Read path | S3 and CloudFront in a separate region | Serves the status page and its data files |
A visitor's request only ever touches the read path. If the API, the database and trigger.dev are all down at once, the last published page still serves.
What works today
Galena is in active development and has no release yet. Working now:
- Sign-in with email and password, TOTP two-factor at sign-in, first-run owner setup, and roles.
- Components and groups, HTTP monitors with per-region results, and detection across regions.
- Incidents with updates and impact, and maintenance windows that start and finish on their own.
- The static status page with a 90-day history per component, incident pages, RSS and Atom feeds, a status badge and state favicons.
- Email subscribers, Slack incoming webhooks and signed outgoing webhooks.
Drafting incidents automatically from monitors, approvals in Slack, alert ingest, a Statuspage-compatible API, settings, invites and GitHub sign-in are next. See the roadmap.
License
Galena is open source under the Apache License 2.0.