PETROS
Designs your platform, checks it continually and writes every change for you to approve.
Works at the level of a solutions architect: it designs the platform and the patterns your services follow, checks your cloud, infrastructure code and pipelines for security, cost and reliability work, and writes each change as a pull request for your team to approve. Building a service goes to Tiphys as a work order. It never changes anything without a person’s yes.
app.petros.intelligant.ai
- Name
- Petros, Greek for rock
- Works with
- AWS · Kubernetes · Terraform
- Changes
- Pull requests only
- Approval
- Always a person
- Learns
- From what your team decides and what each change did
The design, then the work, in one queue.
Design
The pattern for a new service or tool, and why, from your standards and its reference patterns, before anything is built.
Security
Open ports, broad IAM policies, unencrypted storage and drift from your baseline, each explained in plain language.
Cost
Idle and oversized resources, with what each change would save and what it would affect.
Reliability
Missing limits, single points of failure and risky configuration, found before they page someone.
The change, written for your team to approve.
What PETROS connects to.
Available means built, tested and supported today. Beta works, with the gap noted. Coming is planned and not built yet; ask us if you need it.
Clusters
- KubernetesAvailable
- Amazon EKSAvailable
- Amazon ECSAvailable
- Google Kubernetes EngineAvailable
- Azure Kubernetes ServiceAvailable
- Private Kubernetes clustersthrough a connectorComing
Cloud
- AWSAvailable
- Google CloudAvailable
- AzureAvailable
Code
- GitHubAvailable
- GitLabAvailable
- JenkinsComing
Infrastructure as code
- TerraformAvailable
- Terraform Cloud and EnterpriseAvailable
- AWS CloudFormationAvailable
- Pulumi CloudAvailable
- Pulumi (self-managed state)reads the checkpoint in its bucketBeta
- InfracostAvailable
Delivery
- Argo CDAvailable
- FluxAvailable
- HelmAvailable
Metrics
- DatadogAvailable
- PrometheusAvailable
- Amazon CloudWatch alarmsAvailable
- Grafanaas a Prometheus endpointBeta
Alerts and on-call
- PagerDutyAvailable
Chat
- SlackAvailable
- DiscordAvailable
- WebhooksAvailable
- Microsoft TeamsComing
Tickets
- JiraJira Service Management tooAvailable
- ServiceNowAvailable
- LinearComing
Secrets
- VaultComing
Sign-in
- Single sign-on (OpenID Connect)Available
- Single sign-on (SAML)Coming
The engineer approves the pull request
Findings arrive explained, with the fix already written as code. Nothing merges, applies or restarts until someone says yes.
Connect
Give PETROS read access to your AWS accounts, clusters and Terraform repositories.
Review the finding
Each finding says what is wrong, why it matters and what the fix changes.
Approve the pull request
The fix is already written as code. Merge it, change it or close it.
See PETROS on your own work.
A person on the team reads every request and replies by email.