Varqeni Help

Read your services under Service

Updated 11 November 2026

Service Operations probes each service from outside, the way your users reach it, and reads scheduled jobs and deploys from Kubernetes, read-only.

AvailabilityAvailable to everyone with early access to Varqeni.

Steps

  • In the sidebar, choose Home, then Service operations, or Service in the switch above the page.
  • Endpoints lists each service with its HTTP status and response time, the share of its requests failing and its traffic where the service counts its own requests or sends OpenTelemetry traces, the services it calls (written in its setup, or seen in its traces), and the Deployment behind it where one is named.
  • Scheduled jobs lists each CronJob with its schedule and when it last succeeded.
  • Changed in the last 30 minutes lists the Deployments whose spec changed: the first suspects when a service starts failing.
  • Findings name the cause: a recent deploy broke a service, a workload is hung, a service does not answer or answers with errors, more of a service's requests fail than its threshold, a service is slow, or a scheduled job failed or missed its schedule. When a service fails because a service it calls is failing, Varqeni names the cause once, at the service where it started, and lists the caller as a consequence.

What you should see: For a bad deploy Varqeni recommends rolling back; for a hung workload, a rolling restart. An approver runs either from the incident, and Varqeni checks the service answers afterwards. A release that fails only some of its requests is caught from the error rate even when a single check passes, and before and after a fix Varqeni sends a run of 30 requests, not one. In the lab the engine named the cause in 42 of 42 blind trials, including a broken release that only some requests hit and one that a calling service passed on, and the fixes worked in 32 of 32 closed-loop trials, with the wrong fix never sent.

What it can and cannot do

Can

  • Show whether each service answers and how fast, scheduled jobs and recent deploys, and name the cause when one breaks.
  • Roll back a bad deploy or restart a hung workload after a named approval, then check the service answers.

Cannot

  • Read request error rates from your ingress yet, or fix a service whose Deployment it is not told about.

Still stuck? Contact support or email info@fortisdatagroup.com.