EvoHub

What your teams declared, beside what actually runs

Catalog keeps your services, their owners and their promises in one place, and checks them against what an agent in each cluster actually sees running, so unowned workloads, drift between environments and missed SLAs show up on their own instead of in an incident review.

What is EvoHub Catalog?

EvoHub Catalog is a service catalog that does not rely on people keeping it up to date. Teams declare their services (owner team, tier, runbook, repository) and a read-only agent in each Kubernetes cluster reports what is actually running. Catalog matches the two, proposes owners for what nobody claimed, shows drift between environments, draws the service map and measures each service's SLA from its Uptime monitors and On-Call alerts.

See what runs, and what nobody owns

The Overview counts every application the agents report, what needs attention and how much of it has an owner. Each number opens the list behind it.

  • Applications grouped the way Kubernetes already groups them: a Helm release, an Argo CD application, or a workload with its Services, ingresses, autoscalers and volumes
  • Health per application (healthy, degraded or down) with the workloads that are not ready and why
  • Applications without an owner, listed until someone chooses their service
  • Recent changes: what the last sweeps saw appear, change or disappear

Ownership that keeps itself current

Declare a service once; what the agents find is matched to it, or proposed for a person to confirm.

  • Services with an owner team, a tier, a runbook, a repository and docs links
  • Resources join a service by the evohub.io/component annotation, or are proposed in a match queue ordered by confidence
  • Accepting a proposal puts the application and everything in it in that service; rejecting it means the pair is never proposed again
  • An optional RACI per service (responsible, accountable, consulted, informed) beside the owner team

Drift between environments, before it bites

Where what was declared and what is running disagree, the Drift page says so.

  • Services that run in one environment and are missing from another
  • Workloads that run different image versions in staging and production
  • Resources that disappeared since the last sweep
  • Images pinned to latest or to no tag and no digest, which run whatever the registry held when their pods last started

A service map drawn from the cluster

Applications by namespace, coloured by health, owner or tier; open one for its tree from Ingress to Pod.

  • Edges derived from what the cluster says: Services selecting pods, ingresses routing to Services, autoscalers and volumes
  • Who calls whom, as your teams declared it, drawn apart from what was observed
  • Every container image in use, most used first: the list a security advisory asks for
  • Filters live in the address, so a view is a link you can share

Tiers and SLAs, measured rather than promised

Your own tiers say how critical a service is and what that promises; Catalog measures each service against them every month.

  • Tiers you define, each with availability, response and resolution targets; a service can change any target for itself
  • Availability from the Uptime monitors tied to the service
  • Response and resolution from the alerts its On-Call integrations and monitors raised
  • A monthly SLA report per service, for the last twelve months

A read-only agent in your cluster

Nothing is read without an agent you install yourself, with one Helm command and a single-use token from Organization → Agents.

  • Read-only access, and no access to Kubernetes Secrets
  • Secret values and environment variable values never leave your cluster
  • One environment per cluster; the agent reports its heartbeat, and a late agent is flagged
  • Export the whole inventory, every asset with its service, owner, tier, SLA and RACI, as CSV or JSON for an audit

Pricing

You pay for each environment being watched: 11 EvoHub tokens a day, about $5 a month, for every environment an agent or a cloud account reports from, however many clusters or machines are in it. Services, agents, cloud accounts, the resources they discover, the map, drift and SLA reports are free, so describing everything and connecting everything never makes the bill grow. Every account's monthly free tokens apply. See pricing.

Who sees what

Each part follows its own permission: who may read and change the declared services, who may read what the agents observed, and who may manage environments. Agents and their enrolment tokens belong to the organization and are managed beside its API keys.

Frequently asked questions

What is EvoHub Catalog?

EvoHub Catalog is a service catalog that does not rely on people keeping it up to date. Teams declare their services (owner team, tier, runbook, repository) and a read-only agent in each Kubernetes cluster reports what is actually running. Catalog matches the two, proposes owners for what nobody claimed, shows drift between environments, draws the service map and measures each service's SLA from its Uptime monitors and On-Call alerts.

What does the Catalog cost?

You pay for each environment being watched: 11 EvoHub tokens a day, about $5 a month, for every environment an agent or a cloud account reports from, however many clusters or machines are in it. Services, agents, cloud accounts, the resources they discover, the map, drift and SLA reports are free, so describing everything and connecting everything never makes the bill grow. Every account's monthly free tokens apply.

What does the agent read, and what never leaves my cluster?

The agent runs with read-only access to the Kubernetes API and no access to Secrets. It reports workloads, pods, Services, ingresses, autoscalers, volumes and the images they run. Secret values and environment variable values are never sent.

How are resources matched to services?

A resource with the evohub.io/component annotation joins that service directly. Everything else is grouped into applications and proposed for a person to accept or reject; confidence orders the queue, and nothing is accepted on its own.

How is a service's SLA measured?

Availability comes from the Uptime monitors tied to the service; response and resolution from the alerts raised by its On-Call integrations and monitors. Each month is compared with the targets of the service's tier, or the service's own changes to them.

Can I export the catalog for an audit?

Yes. The Overview exports every asset with its service, owner team, tier, SLA targets and RACI, as CSV for a person or JSON for a tool.

Start free — 200 EvoHub tokens every month. See pricing.

Last updated: · EvoHub is built and operated by EvoSync Teknoloji A.Ş.