OpsKnight 1.4.0 · Apache-2.0 · you run it

The 2am page should live on your servers — not in a SaaS you rent per person.

When production breaks, OpsKnight messages whoever is on call, opens a Slack room, and gives customers a status page. Afterward you write what happened. One product. Your machines.

Built for teams that already run Postgres and would rather operate a pager than pay per seat. Not a hosted cloud, and not native voice calls.

Install GitHub Live status

License: Apache-2.0
Parsers: 27
Deploy: Compose / Helm
Hosting: Your VPC

Command center · OpsKnight 1.4.0

OpsKnight command center on a real install — open incidents, on-call, and MTTA/MTTR

The night, in order

Alert, page, talk, write it down.

The same sequence most teams already run. OpsKnight is the place that sequence lives — on your network.

01
Something fires
Monitoring tools send a webhook. OpsKnight understands 27 of those out of the box, plus generic JSON.
02
Noise is grouped
Related alerts become one incident, so one person is not messaged for every downstream symptom.
03
Someone is paged
The active rotation gets SMS, email, Slack, WhatsApp, or a push notification. If they miss it, the policy moves on.
04
The team gathers
A Slack channel is opened for that incident. Optional video link. Work stays on one timeline.
05
You close it out
Resolve, update the status page, and write what happened — with follow-up tasks if you need them.

The product

Screens from a real install.

Same sequence as the night: list, schedule, Slack, policy, service. Status is one page. Ack from a phone is the site on the home screen — no extra app-store product.

One list for what is open.

Incoming webhooks become incidents. You acknowledge, assign, and resolve. Related alerts can share a fingerprint so one person is not paged for every downstream symptom.

Open, acknowledged, and resolved in one place
Filter by service, urgency, and who is assigned
Write the postmortem from the timeline — OpsKnight does not invent a report
Ack from a phone by installing the site; same login as desktop

What you get

The night, in one install.

Apache-2.0. You host it. You still pay compute, SMS, and Slack. There is no OpsKnight cloud and no per-seat software fee.

01
Incident command
Open, ack, assign, resolve. Related alerts can share a fingerprint so one person is not paged for every downstream symptom.
02
On-call schedules
Rotations, timezones, and last-minute swaps. The active layer is who gets the first message.
03
Escalation and paging
Email, SMS (Twilio or SNS), Slack, WhatsApp, push, or a webhook. If they miss it, the policy moves on. No native voice calls.
04
Customer status page
Public, restricted, or private. Optional custom domain and email subscribe. One page per install.
See how it ships →
05
Analytics and SLA
MTTA, MTTR, and compliance on your Postgres. You read the numbers; OpsKnight does not invent a postmortem.
06
Ack from a phone
Install the site on the home screen. Same login as desktop. Not a separate app-store product.

Install

Run it where your systems already live.

Compose, Docker, Helm, or Kustomize. Postgres, NEXTAUTH_SECRET, and ENCRYPTION_KEY are required. There is no hosted cloud.

0.5 vCPU512MB RAM

curl -sL https://raw.githubusercontent.com/opsknight-labs/OpsKnight/main/docker-compose.yml > docker-compose.yml
docker compose up -d

OpsKnight requires PostgreSQL, NEXTAUTH_SECRET, and ENCRYPTION_KEY. The bundled Docker Compose configuration starts both PostgreSQL and OpsKnight automatically.

First-Boot SecretsRequired

NEXTAUTH_SECRET

Generating...

ENCRYPTION_KEY

Generating...

PostgreSQL 14+ is required (bundled in Compose).

AES-256 & NextAuth JWT

Need the longer walkthrough? Install page · Docs · Helm charts

Integrations