Operate
You sleep.
Blayde keeps watch.
Every app Blayde builds is checked from outside every five minutes and its alarms are read. When something breaks, everything related joins one incident and you're paged once, on Slack or Teams.
- Checked every five minutes
- One page per problem, rated P1 to P4
- Risky changes heard in seconds
Watch Live
An operations centre on everything it builds
Every app Blayde deploys is checked from outside every five minutes and its alarms are read. Trouble opens an incident, rated P1 to P4, and everything related joins it, so one problem pages you once, on Slack or Teams, and again when it's resolved.
Audited accounts are watched too: re-audited once a day while your approval holds, and a new serious finding opens an incident.
Every five minutes is 2,016 checks a week. The incident and page are an example.
Live events Live
It hears about trouble in seconds
Turn on live events with one click in your AWS account. Alarms, crashed containers, AWS outages and risky changes reach Blayde the moment they happen: a port opened to the internet, access changed, the audit trail switched off.
Event details only, never your data.
An example night
You slept eight hours. Blayde didn't.
- All 4 of your apps healthyChecked from outside, every five minutes.
- Checkout errors jumped to one request in tenThe alarm Blayde built into your API reached it in seconds. Incident opened, P2, and you were paged once on Slack.
- Containers restarting · exit 137Out-of-memory kills on the checkout service. Joined to the same incident, so no second page.
- Root cause: a memory limit set too low ComingYesterday's deploy halved the limit; the 2 a.m. invoice run hit it. One cause that explains every symptom.
- A pull request is waiting for you ComingPuts the limit back, adds an alarm at 80% of it, and a test that reproduces the crash.
- Port 22 opened to the whole internet, by handSomeone edited a security group outside Blayde. Joined to the open incident with who did it and when.
- Morning summary ComingWhat broke, why, who was affected and for how long, in plain words.
An illustrative night. Steps marked Coming are being built now.
Self-healing Coming
When an update breaks your app, one tap puts it back
When Blayde's own update breaks an app, it will spot it, work out why, and offer to put back the version that worked. You tap "Fix it", send it to your developer, or say not now. Nothing is fixed without a person saying yes.
In Blayde's own drill, a deliberately broken release was caught by the five-minute check and, after one tap, the app was answering normally again about a minute later.
Scale Coming
“We expect 10x traffic in March”
Blayde will read how your system is doing, plan the capacity, caching or database change with its new monthly cost and a load test, and send it through the same approval as everything else.
Operating is one part of the loop
Blayde also builds and secures what it runs. See all of it from one place.