The incident lands in Slack. So does the diagnosis.

An alert from your monitoring tool posts to your enrolled channel. Echelar reads the repository at its exact revision, correlates the alert with recent changes, and posts the cause and a proposed fix in the thread. Code changes wait for a go-ahead.

  1. Alert 09:41
  2. Read-only investigation
  3. Diagnosis in the thread
  4. Go-ahead from a person
  5. Pull request 842
Replay your last incident with EchelarWatch the demonstration

Thirty minutes with the founder. One repository, read-only, is all it needs.

Incident response before Echelar

Paged at 2am with a stack trace and no context

The alert says what broke. Nobody is around to say why.

The one person who knows that subsystem is asleep

Every incident becomes a search for the same two people.

The postmortem lists last quarter's cause

The fix was known. It never became a rule.

Demonstration

Investigate and fix production errors from Slack.

An alert posted to Slack starts the investigation. Echelar reads the alert, correlates it with recent changes, reproduces the failure in its sandbox, explains the cause, waits for approval, returns the pull request to the same thread, and hands it to the PR Reviewer.

Demonstration — authored data

Sequence description

  1. An alert from the monitoring tool is posted in the enrolled production-errors Slack channel.
  2. Echelar opens a repository-scoped investigation in the same Slack thread.
  3. It reads the alert: the error class, the release, and the affected code path.
  4. It correlates the alert with recent commits to the retry path at the repository's exact revision.
  5. It reproduces the failure in an isolated sandbox. Nothing touches production.
  6. Echelar posts its diagnosis and asks a person on the team to confirm before any code change.
  7. After explicit confirmation, Echelar applies the patch and executes repository checks.
  8. A pull request is linked in the Slack thread and handed to the PR Reviewer.
  9. The PR Reviewer checks pull request 842 against the repository architecture, guardrails, and required checks. A person on the team decides what ships.

Demonstration on a fictional Acme repository — authored data and timings.

How Echelar investigates an incident from Slack

Recommendation first
The investigation is read-only: the alert itself, the repository at its exact revision, recent commits, and what it can reproduce in its own sandbox. It posts its diagnosis and a proposed patch.
The evidence is in the thread
Every tool call — code search, git history, a test in the sandbox — appears as it happens. On-call reads what Echelar read, in the channel where the alert landed.
After the go-ahead, a pull request
The patch and its regression test land on a branch as a pull request. Your CI judges it; your team merges it.
The cause becomes a rule
When your team corrects the same thing twice, Echelar drafts the rule as a versioned Skill, evidence cited, for a person to enable. Next quarter's postmortem does not list this cause.

Incident response, before and after Echelar

Without EchelarWith Echelar
Who gets pagedWhoever is on call, then whoever actually knows the code.On-call reads the cause and a recommendation before waking anyone else.
Reconstructing contextTabs across the alert, the dashboards, GitHub, and the deploy log.One investigation, grounded in the exact repository revision, posted where the alert landed.
How the fix reaches reviewSomeone patches it at 2am and hopes review catches the rest.A scoped pull request with the regression test, waiting for review.

Same job, on your repository.

Replay your last incident with Echelar

Incident response FAQ

Which monitoring tools does this work with?

Any tool that posts to Slack. The trigger is a message in an enrolled channel, whatever posts it. Echelar reads the alert text and investigates from the repository, its history, and its own sandbox; it does not pull logs or traces from your observability tools.

What can the investigation touch?

It reads the alert, the repository at its exact revision, and recent commits, and it can run the code in an isolated sandbox. It cannot change production or your default branch. A fix arrives as a pull request on a branch.

What if the diagnosis is wrong?

Then nothing has happened yet. Reply in the thread with what you know and it keeps going, or stops. Every step is visible, so you can see exactly where it went sideways.

How long does setup take?

Install the GitHub App, enroll a Slack channel, point your alert source at it. The first real alert is the test.

What does it cost?

Echelar is launching now. Early teams work directly with the founder and agree pricing together.

Bring the real one.

Bring an alert you already fixed. Watch Echelar find the cause on your repository, and stop where you would want it to.

Replay your last incident with Echelar

Not ready to book? Read how access and approvals work first.

One team of agents. Four more jobs.